Dawn–Dusk Asymmetry of the Io Plasma Torus Derived from Io’s Auroral Footprints Observed by Juno-UVS - IOPscience

Dawn–Dusk Asymmetry of the Io Plasma Torus Derived from Io’s Auroral Footprints Observed by Juno-UVS - IOPscience
The American Astronomical Society (AAS)
, established in 1899 and based in Washington, DC, is the major organization of professional astronomers in North America. Its membership of about 7,000 individuals also includes physicists, mathematicians, geologists, engineers, and others whose research and educational interests lie within the broad spectrum of subjects comprising contemporary astronomy. The mission of the AAS is to enhance and share humanity's scientific understanding of the universe.
The following article is
Open access
Dawn–Dusk Asymmetry of the Io Plasma Torus Derived from Io’s Auroral Footprints Observed by Juno-UVS
Shinnosuke Satoh
Vincent Hue
Fuminori Tsuchiya
Shotaro Sakai
Yasumasa Kasaba
Hajime Kita
Masato Kagitani
Alessandro Moirano
Bertrand Bonfond
Hiroaki Misawa
Published 2026 February 9
© 2026. The Author(s). Published by the American Astronomical Society.
The Planetary Science Journal
Volume 7
Number 2
Citation
Shinnosuke Satoh
et al
2026
Planet. Sci. J.
34
DOI
10.3847/PSJ/ae3678
Article
PDF
Article
ePub
You need an eReader or compatible software to experience
the benefits of the ePub3 file format
Shinnosuke Satoh
AFFILIATIONS
Planetary Plasma and Atmospheric Research Center, Graduate School of Science, Tohoku University, Miyagi, Japan; shinnosuke.satoh@pparc.gp.tohoku.ac.jp
Aix-Marseille Université, CNRS, CNES, Institut Origines, LAM, Marseille, France
EMAIL
shinnosuke.satoh@pparc.gp.tohoku.ac.jp
Vincent Hue
AFFILIATIONS
Aix-Marseille Université, CNRS, CNES, Institut Origines, LAM, Marseille, France
Fuminori Tsuchiya
AFFILIATIONS
Planetary Plasma and Atmospheric Research Center, Graduate School of Science, Tohoku University, Miyagi, Japan; shinnosuke.satoh@pparc.gp.tohoku.ac.jp
Shotaro Sakai
AFFILIATIONS
Faculty of Environment and Information Studies, Keio University, Kanagawa, Japan
Yasumasa Kasaba
AFFILIATIONS
Planetary Plasma and Atmospheric Research Center, Graduate School of Science, Tohoku University, Miyagi, Japan; shinnosuke.satoh@pparc.gp.tohoku.ac.jp
Hajime Kita
AFFILIATIONS
Department of Information and Communication Engineering, Tohoku Institute of Technology, Miyagi, Japan
Masato Kagitani
AFFILIATIONS
Planetary Plasma and Atmospheric Research Center, Graduate School of Science, Tohoku University, Miyagi, Japan; shinnosuke.satoh@pparc.gp.tohoku.ac.jp
Alessandro Moirano
AFFILIATIONS
Laboratory for Planetary and Atmospheric Physics, Space Sciences, Technologies, and Astrophysical Research Institute, University of Liége, Liége, Belgium
Institute for Space Astrophysics and Planetology, National Institute for Astrophysics (INAF-IAPS), Rome, Italy
Bertrand Bonfond
AFFILIATIONS
Laboratory for Planetary and Atmospheric Physics, Space Sciences, Technologies, and Astrophysical Research Institute, University of Liége, Liége, Belgium
Hiroaki Misawa
AFFILIATIONS
Planetary Plasma and Atmospheric Research Center, Graduate School of Science, Tohoku University, Miyagi, Japan; shinnosuke.satoh@pparc.gp.tohoku.ac.jp
Rikuto Yasuda
AFFILIATIONS
Planetary Plasma and Atmospheric Research Center, Graduate School of Science, Tohoku University, Miyagi, Japan; shinnosuke.satoh@pparc.gp.tohoku.ac.jp
LIRA, Observatoire de Paris, Université PSL, Paris, France
Article metrics
676
Total downloads
Video abstract views
Share this article
Dates
Received
2025 October 7
Revised
2026 January 8
Accepted
2026 January 8
Published
2026 February 9
Unified Astronomy Thesaurus concepts
Jupiter
Io
Jovian satellites
Journal RSS
Sign up for new issue notifications
2632-3338/7/2/34
Abstract
The Io plasma torus (IPT) is a dense plasma cloud that corotates around Jupiter near Io’s orbit. A dawn-to-dusk electric field shifts the IPT dawnward. The shift orientation has been previously measured, but the observing geometries have limited our understanding of the spatial extent of the dawn-to-dusk electric field in the day–night direction. This study probes the dawn-to-dusk electric field using Io’s auroral footprints. The position of the Io footprint corresponds to the travel time of the Alfvén waves from Io, which depends on the local plasma mass density along the magnetic flux tube connected to the moon. By modeling the Alfvén wave propagation, we retrieved the ion mass density and temperature at Io’s orbit from the footprint positions measured by the Juno Ultraviolet Spectrograph from 2016 to 2022, but the two parameters are degenerate. We found that the flux tube mass content (FTMC)—the total mass integrated along the magnetic field line—is a more robust proxy for variability in the plasma conditions at Io’s orbit. The deduced Io-FTMC is correlated with Io’s local time (LT) in the magnetosphere. The peak Io-FTMC is found at 03:42 ± 00:53 LT, suggesting the dawn-to-dusk electric field is oriented toward a predusk direction. The estimated dawn-to-dusk electric field is 3.0–20.0 mV m
−1
, with an upper limit exceeding the previously observed values. This may reflect temporal variations in the solar wind dynamic pressure and in the local plasma conditions driven by Io’s volcanic activity over the Juno era.
Export citation and abstract
BibTeX
RIS
Previous
article in issue
Next
article in issue
NASA ADS Record
About Related Links
Original content from this work may be used under the terms of the
Creative Commons Attribution 4.0 licence
. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
1. Introduction
Io is the major plasma source in the Jovian magnetosphere, originating from its volcanic activity (reviewed by F. Bagenal & V. Dols
2020
). Outgassed SO
gets dissociated and ionized by electron impact and charge exchange, producing ions that are immediately picked up by the rotating magnetic field. The fast corotation is the root cause for the large centrifugal force, which confines these ions around the centrifugal equator, i.e., the farthest points along the magnetic field lines from the spin axis, forming the Io plasma torus (IPT) around the orbit of Io at a distance of 5.9
from Jupiter (1
= 71,492 km = Jupiter radius). Heavy ions in an outer region of the IPT, called the “warm” torus, are transported radially outward by interchange instability, forming a plasma disk that extends to the orbits of Europa (at 9.4
) and Ganymede (at 15.0
) and beyond.
Spatiotemporal variability of the plasma environment around Io has several origins. The periodic change of the plasma environment due to the tilted magnetic dipole axis with respect to the rotation axis is significant as Io wiggles vertically within the IPT by approximately ±0.7
, which is comparable to the ion scale height of the torus ∼1
(E. G. Nerney et al.
2025
). Transient density increases due to Io’s volcanic activity are also prominent. In the volcanic event in early 2015 (M. Yoneda et al.
2015
; R. Koga et al.
2018
), the electron density averaged between the dawn and dusk ansae at ∼6
increased from 1790 cm
−3
in 2015 January to 2400–2600 cm
−3
in 2015 February (K. Yoshioka et al.
2018
; R. Hikida et al.
2020
). Moreover, due to a dawn-to-dusk electric field, the inner regions with larger plasma density are closer to Io’s orbit in the morning sector of the IPT but farther away in the afternoon sector.
The IPT corotates almost rigidly with Jupiter due to the strong corotation electric field. A dawn-to-dusk convection field is superimposed over the corotation field, shifting the IPT toward the dawn sector. The dawnward shift results in adiabatic heating of electrons in the dusk sector, which brightens the extreme ultraviolet (EUV) emissions at dusk (B. R. Sandel & A. L. Broadfoot
1982
; D. D. Barbosa & M. G. Kivelson
1983
). The offset toward dawn, 0.1–0.2
(e.g., N. M. Schneider & J. T. Trauger
1995
), has been measured by observations of the dense and bright region called the “ribbon.” The ribbon corresponds to the peak flux tube content of the torus (F. Bagenal
1994
) and is located at a distance of, i.e., 5.77
(dawn) and 5.58
(dusk) on average (H. Kondo et al.
2024
), separating the inner cold and outer warm tori. The brightness and vertical scale height of ribbon have been observed from the ground using the bright S
emissions at 673.1 and 671.6 nm with the coronagraphic techniques, providing a long-term monitor of the IPT over several oppositions (e.g., H. Nozawa et al.
2004
; J. P. Morgenthaler et al.
2024
). The magnitude of the dawn-to-dusk electric field has been estimated from the radial position of the ribbon as well as the EUV brightness ratio between the dawn and dusk ansae (D. D. Barbosa & M. G. Kivelson
1983
; W. H. Ip & C. K. Goertz
1983
1984
), assuming a uniform electric field completely aligned with the true dawn–dusk. It varies with the dynamic pressure of the solar wind at the front of the Jovian magnetosphere, with 3.8 mV m
−1
on average and 8.6 mV m
−1
at maximum (G. Murakami et al.
2016
).
Several studies reported that the dawn-to-dusk electric field is not aligned with the true dawn–dusk. W. H. Smyth et al. (
2011
) inferred the orientation of the dawn-to-dusk electric field toward the post dawn–dusk of the magnetosphere (07:20 LT), based on two separate spacecraft observations in the afternoon sector (by Voyager 1 in 1979 and Galileo in 1995) and the ground-based observations in 1991. In contrast, C. Schmidt et al. (
2018
) reported an orientation of 10
± 4
earlier than the true dawn (05:20±00:20 LT) based on the ground-based observations in the 2010s, utilizing geometric parallax over years, but the viewing angle was still restricted to ±11
(44 minutes) from the true dawn–dusk. Restrictions imposed by spacecraft trajectories and line of sight integration in the ground-based observations have made it challenging to obtain a dataset that covers the full 24 hr local time (LT). This limitation constrains our understanding of the structure of the electric field, especially the spatial extent in the noon–midnight direction.
The Juno mission (S. J. Bolton et al.
2017
) has been exploring the Jovian magnetosphere, conducting more than 75 close flybys at the polar regions of Jupiter. Juno’s polar orbits are optimized for latitudinal scans of the Jovian magnetosphere. Using Juno data, properties of the IPT have been studied. For example, the electron density was determined by the Waves instrument at a magnetic shell of
= 5, including low latitudes (W. S. Kurth et al.
2025
). Juno’s radio occultation observations revealed that, depending on the observation epoch, the IPT exhibits a different latitudinal offset at similar longitudes, which might be due to the dawn–dusk radial asymmetry of the IPT (A. Moirano et al.
2021a
). However, the orbital inclination, by design, constrains observations inside the IPT, especially near the equatorial plane, across the LT. Even with the two close flybys at Io on 2023 December 30 and 2024 February 3, in situ measurements remain insufficient to characterize the dawn-to-dusk electric field in the inner magnetosphere.
Juno’s infrared (IR) and ultraviolet (UV) observations in the polar region have provided unprecedented datasets of the aurorae on Jupiter with viewing geometries that are not accessible from the ground or space telescopes. One of the outstanding auroral features in Jupiter is the satellite auroral footprints, induced by Alfvén waves that are launched at the satellite through the electromagnetic interaction with the Jovian magnetosphere (J. E. P. Connerney et al.
1993
; J. T. Clarke et al.
1996
; R. Prangé et al.
1996
). The satellite auroral footprints have been observed by Juno in a wide range of both the System III longitudes and the LTs (A. Mura et al.
2018
; V. Hue et al.
2019
2022
2023
; A. Moirano et al.
2021b
2024a
; J. Rabia et al.
2025
).
The Alfvénic perturbations accelerate auroral electrons both toward and away from Jupiter at high latitudes, which results in multiple auroral spots in Jupiter’s atmosphere. The main Alfvén wing (MAW) spot is associated with the initial Alfvén waves. The reflected Alfvén wing spot is caused by Alfvén waves that undergo more than one reflection in the latitudinal density gradient of the plasma torus/disk. The transhemispheric electron beam (TEB) spot is induced by electrons accelerated in the opposite hemisphere (B. Bonfond et al.
2008
; S. L. G. Hess et al.
2010
2011
; V. Hue et al.
2022
).
The satellite auroral footprints occur azimuthally downstream with respect to the magnetically mapped position of the satellite on Jupiter because (i) the Alfvén wave travels along a magnetic field line at a finite speed and (ii) the corotation velocity of the magnetospheric plasma is faster than the orbital speed of the Galilean satellites. The azimuthal separation between the auroral footprint and the magnetically mapped position of the satellite is called the lead angle. The lead angle is proportional to the travel time of the Alfvén wave from the satellite to the auroral footprint (V. Hue et al.
2023
; S. Schlegel & J. Saur
2023
). The Alfvén velocity
depends on the magnetic field magnitude
, the permittivity of free space
, and the local plasma mass density
Most of the Alfvén wave travel time is spent within the IPT or the plasma disk because of the larger mass density and smaller magnetic field magnitude, and hence, the footprint position translates the plasma conditions at the satellite orbit. Recent studies proved that the footprint position (lead angle) is useful to inversely estimate plasma density and temperature and detect temporal changes in these plasma parameters around Io (A. Moirano et al.
2023
2025
; S. Schlegel & J. Saur
2023
), Europa (S. Satoh et al.
2024
), Ganymede (B. Bonfond et al.
2013
), and Callisto (J. Rabia et al.
2025
). The temporal changes found in these previous studies might be related to Io’s volcanic activity, but discussions have not been closed yet.
In addition to the temporal variations, the lead angle of Io’s auroral footprints is also expected to be correlated with the moon’s LT. The distance between Io and the dense ribbon indeed changes as a function of the moon’s LT; when Io is on the dawn side of the magnetosphere, the ribbon is closer to Io’s orbital distance, and the plasma density around Io is expected to be higher than on the dusk side. The scope of this present study is to analyze the LT effect on the plasma parameters at Io’s orbit using the footprint lead angle observed by the Ultraviolet Spectrograph (UVS) on Juno (G. R. Gladstone et al.
2017
). Section
outlines the measurement of the position of Io’s auroral footprints. In Section
, we describe our retrieval procedure in detail: how we model the IPT and compute the lead angle of Io’s auroral footprints to retrieve ion parameters. In Section
, the retrieval results are presented, and we show that a quantity called the flux tube mass contents (FTMC) is a characteristic quantity for the variation of the IPT. We examine the correlation between the FTMC and Io’s LT and estimate the orientation of the IPT shift due to the dawn-to-dusk electric field. In Section
, we reconstruct the equatorial morphology of the IPT using an IPT density model. Based on the reconstructed ribbon positions, we infer the magnitude of the dawn-to-dusk electric field. Finally, Section
summarizes the key findings of this study.
2. Observations of the Io Footprints
The UV auroral footprint positions of Io have been measured and further characterized following a similar methodology as described in V. Hue et al. (
2023
). Figure
illustrates two example images of Io’s auroral footprints captured by the Juno-UVS instrument. Because the UVS instrument operates in a scanning mode, constrained by Juno’s spin-stabilized nature, the footprint positions were measured from consecutive sets of two spin-average UVS observations (one spin takes 30 s), to limit the smearing of the auroral footprint due to the footprint moving in Jupiter’s frame.
Figure 1.
South polar projection of two spin-averaged Io auroral footprint observations, captured by the Juno-UVS instrument on perijoves 28 (top panel) and 31 (bottom panel). The nadir-looking time for the spin recorded on perijove 28 is 25 July 2020 07:13:51 and 07:14:21. For the perijove 31 data, the corresponding times are 2020 December 30 23:10:01 and 23:10:31. The red cross represents the Juno subspacecraft point at the median recorded time.
Download figure:
Standard image
High-resolution image
UVS is equipped with a scan mirror, which allows pointing up to ±30
away from Juno’s spin plane. Over a single perijove flyby, a range of scan mirror positions is used to point at the various auroral features of interest. The effective exposure time on a point source per Juno 30 s spin depends on (i) the scan mirror position at which it operated, and (ii) whether that point source is observed through UVS’s 0
2-large wide slit, or its 0
025-large narrow slit (T. K. Greathouse et al.
2013
). The typical exposure time per spin at the nominal mirror position (
= 0
from the spin plane) is therefore 30 s × 0
2/360
∼ 17 ms when observed through the wide slit, and 30 s × 0
025/360
∼ 2 ms when observed through the narrow slit. This exposure time scales at 1/
), where
is the mirror position angle with respect to Juno’s spin plane.
The uncertainty in the measured position is defined as the quadratic combination of the uncertainty due to (i) the point-spread function of the UVS instrument (0
1 on the sky, as established by T. K. Greathouse et al. (
2013
)) and (ii) the uncertainty of the projected scale height
of the footprint vertical emission curtain. One improvement compared to the earlier work of V. Hue et al. (
2023
) is the inclusion of the footprint’s characteristic length (i.e., in the magnetic azimuthal direction) and width (i.e., in the radial direction) in the position determination. In that regard, we follow the work of A. Moirano et al. (
2024a
) on the IR auroral footprint. The resulting uncertainty can then be written as:
where
is the squared longitudinal (
) and latitudinal (
) uncertainty of the footprint position. The first two terms of the right-hand side of Equation (
) correspond to the uncertainty factor estimated in the V. Hue et al. (
2023
) study. The third term corresponds to the projected length and width of the footprint on the Jovian longitudinal and latitudinal grid, respectively.
The assumed spot length is 438 ± 156 km in the longitudinal (azimuthal) direction, and a spot width of 154 ± 16 km in the transversal direction, as measured by A. Moirano et al. (
2024a
) in the IR using Juno-JIRAM data (see their Table 1). Previous measurements of the Io spot width and length were also provided by B. Bonfond (
2010
) using the Hubble Space Telescope in the UV, leading to an upper limit of 200 km for the Io spot width, and 850 ± 260 km for the longitudinal length. Because the IR and UV spot width/length measurements agree reasonably well, and because the Juno data is generally less affected by projection effects, we used the A. Moirano et al. (
2024a
) values for the width/length spot sizes. This more realistic error estimation leads to increased uncertainties in the footprint position determination, compared to the work of V. Hue et al. (
2023
), by about 20%–30% in the measured longitudinal spot position and less than 10% in the latitudinal spot position.
Using the size of the spot as an additional source of uncertainty results in an overestimation of the footprint center-location uncertainty by a factor of ∼2 in the northern hemisphere and ∼3 in the southern hemisphere. This choice provides a conservative estimate of the spot’s latitude and longitude uncertainties, given that the spot locations were manually identified and that several other important sources of uncertainty must also be considered. For instance, uncertainties in the magnetically mapped footprint track may still amount to several hundreds of kilometers (e.g., J. C. Gérard et al.
2019
), and in some cases up to ∼1000 km (A. Moirano et al.
2024b
). Furthermore, variability in the altitude of the UV footprint emission has also been highlighted by analysis of the limited UVS observations acquired at high emission angles (A. Moirano et al.
2024b
). The present work focuses on lower emission angle measurements from UVS, and our conservative choice in position uncertainty is intended to account for additional systematic effects not otherwise represented in the instrumental or geometrical terms.
This study also deals with the extracted position of the TEB footprint spot as seen by Juno-UVS. The identification of the TEB was only made possible when it led the MAW spot, and not when the TEB is mixed within the MAW spot and tail emission downstream of the MAW spot. The location of the Io footprint aurora is measured in the UVS images of Jupiter’s polar region at a given time
and then magnetically mapped onto Io’s orbital plane by the internal field model JRM33 (J. E. P. Connerney et al.
2022
) and current sheet model Con2020 (J. E. P. Connerney et al.
2020
). The System III longitude of Io,
, is determined at the same time
by the ephemerides. We then measure the equatorial lead angle
, the longitudinal separation between Io and the magnetically mapped position of the auroral footprint on the orbital plane.
The measured equatorial lead angle of Io’s MAW and TEB footprints is shown in Figure
. The MAW lead angle is correlated with the System III longitude of Io because the Alfvén wave travel time depends mainly on the plasma mass density. As Io migrates latitudinally with respect to the IPT with the System III longitude, the plasma density around Io changes accordingly. The dashed lines in Figure
(a) are the Fourier fit curves,
and
, that exhibit a typical correlation between the MAW footprint lead angle and the System III longitude of Io (
):
Figure 2.
Measured equatorial lead angle of the MAW (panel (a)) and TEB (panel (b)) footprints as a function of Io’s System III (SIII) longitude (
). The dashed lines correspond to the Fourier fit curves,
and
(see Equations (
) and (
)).
Download figure:
Standard image
High-resolution image
The same Fourier fit curves are also shown in Figure
(b) for the TEB spots. The TEB footprint is generated by the electron beam that originates from the MAW spot in the other hemisphere. The TEB lead angle is therefore correlated with the Alfvén wave travel time between Io and the MAW spot in the other hemisphere; for example, the northern TEB lead angle mostly depends on the Alfvén wave travel time between Io and the southern MAW lead angle,
. Hence, the correlation with the System III longitude of the TEB is inverse to that of the MAW footprint.
3. Retrieval of the Ion Parameters
3.1. Data Selection
In each orbit, Juno performs a close flyby of Jupiter in the northern polar region first and flies toward the south. Juno is a spin-stabilized spacecraft, and UVS operates in scanning mode, controlled by the spacecraft’s 30 second-spin period. The satellite auroral footprints are therefore not always in the field of view. Since the retrieval procedure described in this section takes three fitting parameters described in Section
3.2
, more than three data points are required. Due to this requirement, perijove data with fewer than four data points of the footprint aurora are eliminated from the analysis. When there are four or more UVS observation data points available in either the northern or southern hemisphere, we analyze the data from each hemisphere separately to improve the temporal resolution of the analysis. After applying the selection criteria, we have 47 data subsets for the Io footprints out of the whole dataset from 43 perijoves.
3.2. Computation of the Equatorial Lead Angle
Unlike previous studies (e.g., S. Satoh et al.
2024
) that used the MAW spots only, we use the equatorial lead angle not only of the MAW spot but also of the TEB spots to add stronger constraints on the parameter retrieval. Here, we describe how we model an IPT density and compute the equatorial lead angle of both MAW and TEB footprints by tracing Alfvén waves in the IPT model with various plasma conditions.
The fast corotation causes a large centrifugal force and confines ions near the centrifugal equator, the farthest location of the magnetic field line from Jupiter. The centrifugal equator lies between the Jovigraphic and magnetic equators, and its Jovigraphic latitude varies with the System III longitude because the magnetic field is tilted with respect to Jupiter’s rotational axis. In this study, we use the Jovian magnetic field community code provided by R. J. Wilson et al. (
2023
) to calculate the magnetic field. We use the Schmidt coefficients up to order and degree 13 for JRM33 and the analytic mode for Con2020. All the input parameters for Con2020 are set to the default values (see Table 2 in R. J. Wilson et al.
2023
). Based on the field model, the centrifugal equator coincides with Io’s orbital plane at the System III longitudes of ∼113
W and ∼281
W.
We model the IPT as a single ion species characterized by atomic mass
, temperature
, and charge
. These three ion parameters, along with the electron temperature (
), are treated as constants along each magnetic field line. The ion number density in the Io flux tube (
) is approximated by a Gaussian function of
, the distance measured from the centrifugal equator along the magnetic field line. The peak density at
= 0 is denoted as
. Latitudinal extent of the ion density profile is quantified by a density scale height of
, which depends on ion temperature (
), electron temperature (
), and the centrifugal force (F. Bagenal
1994
). Accordingly, the Gaussian approximation is described by the following three expressions:
where
is the ion mass density.
= 1.67 × 10
−27
kg is the proton mass, Ω
= 2
/(9 hr 55 minutes 29.71 s) is the angular velocity of Jupiter’s rotation, and
= 1.38 × 10
−23
J K
−1
is Boltzmann’s constant.
Figure
illustrates a mass density profile along the magnetic field line of Io, obtained from Equations (
)–(
). The centrifugal equator,
= 0, is defined by the JRM33 and Con2020 field models. In our retrieval procedure, the fitting parameters (
) and the constants (
) are all treated as azimuthally uniform.
Figure 3.
Illustration of a mass density profile along the magnetic field line of Io. The magnetic field lines (dashed orange lines) are calculated by the JRM33+Con2020 magnetic field model. The footprint positions are indicated in white.
Download figure:
Standard image
High-resolution image
Previous studies established a method to model the IPT density profiles along the magnetic field lines using multispecies diffusion equilibrium (e.g., F. Bagenal
1994
; Y. Mei et al.
1995
; L. P. Dougherty et al.
2017
), which yields separate density distributions for electron and each ion species, in contrast to a Gaussian approximation with a hypothetical single ion species. The diffusive equilibrium method is particularly useful for deriving radial density profiles of individual ion species from ground-based spectroscopic observations of the IPT (e.g., E. G. Nerney et al.
2025
). However, A. Moirano et al. (
2025
) recently reported that the chemical composition of the IPT cannot be retrieved by using only the footprint lead-angle measurements. Since the Alfvén velocity depends not on properties of individual ion species but on the total ion mass density, the ion fractions derived from such analysis are highly degenerate. The Gaussian approximation is less realistic, but, given the degeneracy of ion fractions, it remains sufficient for retrieving spatiotemporal variations of the averaged mass, number density, and temperature in the Io flux tube from the footprint lead angle.
We calculate the local Alfvén velocity:
where
is the magnetic permeability, and
) is the magnetic field magnitude. The Alfvén velocity approaches the speed of light (
) at higher latitudes in the Jovian magnetosphere, where the magnetic field is stronger, and the mass density is much reduced. It therefore needs to be relativistically corrected (P. C. Hinton et al.
2019
):
The one-way travel time of the Alfvén wave from Io to the MAW footprint (
) is obtained by integrating the local Alfvén velocity along the magnetic field line:
When the Alfvén wave approaches Jupiter’s ionosphere, the Alfvén velocity decreases to ∼0.3
at an altitude below 2500 km (K. Saito et al.
2023
), which is more than 2 orders of magnitude faster than that at Io, ∼180 km s
−1
(M. G. Kivelson et al.
2004
). At a speed of 0.3
, the Alfvén travel time within the ionosphere is only ∼0.02 s, whereas the Alfvén travel time from Io to the ionosphere is 4–14 minutes (A. Moirano et al.
2023
). Since the Alfvén travel time within the ionosphere is negligibly small, the ionosphere is not included in the present study.
We compute
numerically with d
= 10 km (Equation (
10
)).
is then converted to the equatorial lead angle of Io’s MAW footprint (
) using a synodic period of Io (
), in the same manner as V. Hue et al. (
2023
) and S. Satoh et al. (
2024
).
We assume that a TEB travels along the magnetic field line from one hemisphere to the other at the speed of light. The TEB travel time,
TEB
, is calculated by tracing the magnetic field line from a MAW footprint to the opposite hemisphere. When Io is at the System III longitude of 113
W, the Io flux tube field line has a length of 14.8
, which is traversed at the speed of light in 3.5 s. The equatorial lead angle of the TEB footprint (
) is then obtained by adding an angular delay due to the TEB travel time to the equatorial lead angle of the conjugate MAW footprint in the opposite hemisphere.
3.3. Fitting to the Observed Equatorial Lead Angle
We chose the three parameters (
) as the fitting parameters. The parameters used in the retrieval procedure are summarized in Table
. Even though we assume a hypothetical single ion species, changes in the atomic mass (
) reflect different ion fractions in the flux tube.
∼ 22 represents the major ion species around 5.9
, S
++
(32 AMU) and O
(16 AMU), whose mixing ratio is ∼0.24 and ∼0.20, respectively (K. Yoshioka et al.
2018
; R. Hikida et al.
2020
). The ion number density (
) and temperature (
) represent bulk values of the IPT. The charge of ion (
) and the electron temperature (
) are both assumed to be constant in the Io flux tube (the sensitivity analysis can be found in the appendix section) at 1.4 and 6.0 eV, respectively (M. G. Kivelson et al.
2004
). The
space sufficiently covers the ranges provided in M. G. Kivelson et al. (
2004
), i.e., 960–2900 cm
−3
, and 20–90 eV, respectively.
Table 1.
Parameter Sets and Ranges Used in the Retrieval Procedure
Parameter
Ion Atomic Mass
Equatorial Number Density
Ion Temperature
Charge of Ion
Electron Temperature
(AMU)
(cm
−3
(eV)
(eV)
Values
20, 22, 24
500–5000
10–1000
1.4 (constant)
6.0 (constant)
Steps
3 (linear)
50 (logarithm scale)
60 (logarithm scale)
Download table as:
ASCII
Typeset image
The footprint lead angles,
and
, are computed in a three-dimensional parameter space of (
) and compared with the observed values in each data subset. We explore the parameter space (Table
) and conduct a grid search for the best-fit parameter set for each data subset. The chi-square value (
) is calculated by
where
is the number of data points in a single data subset,
is the
th observed equatorial lead angle, and
is the measurement uncertainty (described in Section
) mapped onto the equatorial plane.
is the computed equatorial lead angle.
is the position of Io, the launch site of the Alfvén wave of interest; it is not taken at the observation time (
), but at the time when the corresponding Alfvén wave was launched (
). The time difference between
and
is directly inverted from the equatorial lead angle, which leads to the following relation:
where
is the observed lead angle, and
represents either MAW or TEB.
is then obtained from NAIF’s SPICE kernels (C. H. Acton
1996
; C. Acton et al.
2018
) at
We also calculate
to estimate the confidence levels, where
is the minimum of all the chi-square values calculated in the entire parameter space. For a two-dimensional parameter space, Δ
= 2.30, 6.18, and 11.8 correspond to 1
(68.3%), 2
(95.4%), and 3
(99.73%) confidence levels, respectively (W. H. Press et al.
2002
).
4. Results
4.1. The Retrieval of the Ion Parameters
Figure
presents examples of the retrieval results for the PJ3, PJ19, and PJ28-South (PJ28S, hereafter) data subsets. The three data subsets were chosen because they exhibit different density–temperature states: strong degeneracy between the retrieved ion number density and temperature in PJ3 (Figure
(a)), high-density/low-temperature in PJ19 (Figure
(c)), and low-density/high-temperature in PJ28S (Figure
(e)).
Figure 4.
Examples of retrieval results for PJ3 (first column), PJ19 (second column), and PJ28S (third column). The first row shows the heat maps of the reduced chi-square (
per the degree of freedom) as a function of
and
, presenting the goodness of fit. The three contour lines in each panel represent the 1
, 2
, and 3
retrieval confidence levels. The best-fit parameter is indicated as a white dot. The second row shows the best fits to the observed equatorial lead angle as a function of Io’s System III longitude (
). The solid lines and dashed lines represent the best-fit values to the MAW and TEB lead angle, respectively.
Download figure:
Standard image
High-resolution image
The degeneracy between
and
prevents independent evaluation of each parameter. This degeneracy arises because the Alfvén wave travel time (
) is approximately proportional to
(A. Moirano et al.
2023
), but the degree of degeneracy varies among the data subsets. In fact, the degeneracy is weak in other data subsets, and the confidence contours are completely closed (i.e., PJ19 and PJ28S in Figures
(c) and (e), respectively).
We found that the retrieval procedure is not sensitive to variations in the atomic mass
. We therefore adopt a fixed value of
= 22 AMU throughout the remainder of this study (M. G. Kivelson et al.
2004
) since
= 22 well represents the ion composition derived from the plasma diagnostics based on the Hisaki observations (K. Yoshioka et al.
2018
; R. Hikida et al.
2020
). Instead of analyzing each fitting parameter (
) individually, we introduce a composite quantity, the FTMC, in the next section. We show that FTMC provides a robust characterization of the variations in the footprint lead angle in terms of the total plasma mass integrated within the Io flux tube despite this density–temperature degeneracy.
We do not analyze or discuss the fitting parameters from each data subset in this paper because of the degeneracy. However, to benchmark our retrieval procedure, here we assess the retrieval sensitivity to the ion number density and temperature by describing the retrieved parameters from the PJ19 and PJ28S subsets. The PJ3 result is briefly described in the next section. In Figure
, the best-fit parameter is shown as a white dot inside the confidence level contours. The best-fit parameters with the 1
errors are (
) = (
cm
−3
eV) for PJ19 and (
cm
−3
eV) for PJ28S. The best-fit ion number density for both perijoves is within the range of 960–2900 cm
−3
reported by M. G. Kivelson et al. (
2004
). The best-fit ion temperature for PJ19 is within the range 20–90 eV from M. G. Kivelson et al. (
2004
) as well. In contrast, it exceeds the range for PJ28S. However, the lower limit temperature for PJ28S is 82 eV, accompanied by an ion number density of 1696 cm
−3
, and both values agree with the previous values (960–2900 cm
−3
and 20–90 eV from M. G. Kivelson et al. (
2004
)).
The comparison between the two data subsets indicates that PJ19 and PJ28S corresponded to different density–temperature states: high-density/low-temperature in PJ19 and low-density/high-temperature in PJ28S. These results demonstrate that our retrieval procedure yields reliable estimates of ion density and temperature and detects different density–temperature states in the IPT.
In Figures
(b), (d), and (f), the best fits to the measured equatorial lead angle are displayed as a function of Io’s System III longitude (
). We computed the Alfvén wave travel time from Io at every 7
5 of System III longitude using Equations (
)–(
13
) with the best-fit parameters. Variations in the equatorial lead angle in different observation timings are well predicted in Figures
(b), (d), and (f). Since we assume that the mass density profile is uniform at each System III longitude, the periodicity of the equatorial lead angle is only due to the tilted magnetic field morphology.
4.2. FTMCs
The strong degeneracy shown in the previous section prevents us from characterizing the plasma conditions at Io’s orbital distance using the retrieved
and
. To overcome this issue, we introduce a quantity called FTMC and use the FTMC as a characteristic quantity of the plasma condition at Io’s orbital distance, instead of discussing the degenerate fitting parameters. The FTMC is an integrated mass column density along a closed magnetic field line, from the southern to the northern hemispheres (F. Bagenal et al.
2016
), such that:
With Equations ((
)–(
)), the FTMC is a function of all the fitting parameters (
; Equation (
16
)) and approximately proportional to both
and the square root of
As mentioned in the previous section, the degeneracy arises due to
(A. Moirano et al.
2023
).
is not strictly proportional to
(Equation (
10
)), but the idea of using the Io-FTMC originates from the fact that
and
have a similar relationship with both
and
. The comparison between the Io-FTMC heat map and the 3
retrieval confidence level (the first row of Figure
) clearly demonstrates that the degeneracy is correlated with the FTMC in the density–temperature space.
Figure 5.
(Top row) Heat maps of the Io-FTMC as a function of
and
. The Io-FTMC is calculated by Equation (
16
). The heat map is purposely discretized to show the correlation between the FTMC and the retrieval confidence levels. The 3
retrieval confidence levels from Figure
are superposed. (Middle row) Same as the top row, but the region outside the 3
retrieval confidence level is masked. (Bottom row) Histograms of the Io-FTMC obtained within the 3
retrieval confidence level with the same color code as the other two rows. The accompanying box plots indicate the quartiles, minimum, and maximum values.
Download figure:
Standard image
High-resolution image
It may be worth asking whether the FTMC can be selected as the only fitting parameter for the lead-angle fitting in the first place, instead of using all the presented fitting parameters. However, as described in Section
3.2
, the lead-angle fitting requires the local Alfvén velocity (Equations (
) and (
)), which cannot be replaced by any field-line integrated quantities.
Based on Δ
, we extract the grid points within the 3
retrieval confidence level in the density–temperature space (the second row of Figure
), which allows us to obtain the histograms of the Io-FTMC by counting the extracted grid points (the third row of Figure
). We also obtain the box plots that visualize the quartiles, minimum, and maximum values. We conduct the same procedure for each data subset to obtain the time series of the Io-FTMC (Figure
).
Figure 6.
Inferred Io-FTMC as a function of time, accompanied by the corresponding perijove numbers. The blue box plots indicate the quartiles, minimum, and maximum values obtained in the same manner as showcased in Figure
. (b) Io’s LT of the footprint observation.
Download figure:
Standard image
High-resolution image
We here showcase the same three instances from the PJ3, PJ19, and PJ28S data subsets in Figure
. The median value of the Io-FTMC with a quartile range is
kg m
−2
kg m
−2
, and
kg m
−2
, for PJ3, PJ19, and PJ28S, respectively. When strong degeneracy is present in the retrieval result, in the PJ3 subset, for instance, the inferred FTMC has a wide range from minimum to maximum, but the interquartile range is comparable to that of the other data subsets with weaker degeneracy. Additionally, in the strongly degenerate cases,
is found at the upper boundary of either
or
: i.e.,
is located at
= 5000 cm
−3
in PJ3 (Figure
(a)). The exact minimum for the PJ3 subset is found at
= 5395 cm
−3
if the
axis is extended sufficiently, whereas this extension would alter the Io-FTMC median by only ∼0.2% (additional details referring to the sensitivity analysis can be found in the appendix section). Despite the degeneracy between the retrieved ion density and temperature, the FTMC provides a robust characterization of plasma mass in the Io flux tube at each observation timing.
The FTMC (Equation (
16
)) is integrated along the field line connecting the two hemispheres. Some of the data subsets contain measurements of only one type of footprint (MAW or TEB) on one single hemisphere, in which case the Alfvén wave tracing is carried out in only one direction from the moon. Even in such instances, the FTMC can be derived since we assume a plasma density profile symmetrical in the
direction with respect to the centrifugal equator (see Equation (
)), and azimuthally uniform with the System III longitudes.
Figure
(a) shows the inferred Io-FTMC as a function of time. The mean Io-FTMC over time is 1.50 × 10
−8
kg m
−2
, which can be converted to the flux tube mass per unit magnetic flux, 5.71 × 10
−3
kg Wb
−1
. This is well comparable to the in situ observations; it is within the same order of magnitude as ∼2.5 × 10
−3
kg Wb
−1
obtained from the Galileo observations (F. Bagenal et al.
2016
). The variation exhibited in the Io-FTMC (Figure
(a)) should be a result of both (i) temporal changes that may be associated with Io’s volcanic activity and (ii) a systematic change associated with Io’s LT. Due to the dawn-to-dusk electric field, the distance between Io and the dense ribbon changes as a function of the LT, which is expected to alter the plasma parameters at Io periodically with the moon’s LT. To understand the Io-FTMC variation, we characterize the LT effect on the FTMC, which will help future works investigate the temporal changes over the Juno epoch by being compared to both in situ and remote observations of the IPT.
Figure
(b) shows Io’s LT during the UVS footprint observations. Juno was in a 53 day orbit during its prime mission from PJ1 to PJ33, and Io has a 42.5 hr orbital period, which results in a constant synodic period between Juno and Io. After the completion of the prime mission, Juno’s first extended mission started with a close encounter with Ganymede on 2021 June 7, where a gravity assist took place, which reduced Juno’s orbital period to 43 days. The following close flybys at the Galilean moons (at Europa on 2022 September 29, and at Io on 2023 December 30, and 2024 February 3) reduced it further, which consequently has ruffled correlation between the observation dates and corresponding LT of Io over the first extended mission. Thanks to Juno’s repetitive observations over a baseline covering nearly 6 yr, we obtained the Io-FTMC in a full 24 hr LT range. Figure
shows the inferred Io-FTMC as a function of Io’s LT. One can notice a clear correlation between the inferred FTMC and Io’s LT. On the dawn side of the magnetosphere, the ribbon is closer to Io’s orbital distance because the IPT is shifted radially outward, which results in a higher mass content in the Io flux tube than on the dusk side. We fit a cosine function to the inferred FTMC as a first-order Fourier fit using the orthogonal distance regression.
Figure 7.
Inferred Io-FTMC as a function of Io’s LT. The plot is repeated in 24:00–48:00 for clarity. The horizontal error bars correspond to the observing duration of each data subset: data subsets that consist of both the northern and southern observations have larger LT uncertainties. The black solid curve and gray shaded area represent the best-fit cosine function
fit
(Equation (
18
)) and the 3
fitting error, respectively.
Download figure:
Standard image
High-resolution image
As mentioned earlier, the variation of the Io-FTMC may also include the temporal changes in the IPT. Despite this, using the cosine fit (
fit
) allows us to quantify this distinct LT effect on the Io-FTMC associated with the dawn-to-dusk electric field. The obtained parameters are
= 3.22 ± 0.64,
= 3.70 ± 0.88 hr, and
kg m
−2
. The best fit (
fit
) is shown in Figure
along with the corresponding 3
fit error. The phase shift,
, is equivalent to 03:42 LT with an uncertainty of ±53 minutes. This result indicates that a peak value of the Io-FTMC is found in the predawn sector of the magnetosphere. The peak Io-FTMC found in the morning sector is consistent with the fact that the IPT is shifted radially toward the morning sector due to the dawn-to-dusk electric field. Plasma density around Io is expected to be higher when Io is in the morning sector because the ribbon region gets closer to Io due to the dawn-to-dusk electric field. In the next section, we will discuss the spatial structure of the IPT and the dawn-to-dusk electric field from a unique perspective of the Io-FTMC inferred from the footprint lead-angle observations.
5. Discussion
In this section, the best-fit function
fit
is adopted to represent the typical LT effect on the Io-FTMC. Both the orientation and magnitude of the dawn-to-dusk electric field are evaluated using
fit
together with other modeling approaches.
5.1. Pointing of the IPT Shift Due to the Dawn-to-dusk Electric Field
Figure
shows the LT orientation of the radial shift of the IPT inferred in this study. We take the best-fit phase
in
fit
(Equation (
18
)), 03:42 ± 00:53 LT, as a representative orientation of the IPT shift due to the dawn-to-dusk electric field. It is compared with the two previous studies, one of which is C. Schmidt et al. (
2018
), who investigated the orientation of the dawn-to-dusk electric field using a comprehensive set of ground-based observations of the ribbon from 2013 to 2017. They proposed a dawn-to-dusk electric field pointing 10
± 4
earlier than the true dawn, i.e., 05:20 ± 00:20 LT. Despite the limitation of the observation geometry, they exploited parallax to constrain the electric field orientation. The parallax allowed them to sample ±11
in Jovian LT, which is equivalent to only ±0.7 hr from the true dawn–dusk line, and this, therefore, could be a lower limit of the shift orientation with respect to the true dawn–dusk line. However, it still supports our result that the electric field is pointing toward the predawn sector.
Figure 8.
Comparison of the pointing of the radial shift of the IPT in the Jupiter–Sun–Orbit (JSO) coordinate, where Jupiter is located at the origin,
JSO
axis points toward the Sun,
JSO
axis is aligned with Jupiter’s spin axis, and
JSO
axis completes the coordinate.
JSO
< 0 and
JSO
> 0 correspond to the morning and afternoon sectors, respectively. The shaded areas represent the uncertainties of the inferred orientation.
Download figure:
Standard image
High-resolution image
Another study, led by W. H. Smyth et al. (
2011
), proposed a dawn-to-dusk electric field pointing 20
later than the true dawn, i.e., 07:20 LT, based on the Voyager 1 observations, ground-based observations in 1991, and the Galileo J0 observations. A possible interpretation of a variable orientation of the dawn-to-dusk electric field is that the orientation may vary in time: our study and C. Schmidt et al. (
2018
) are based on the observations in Juno and Hisaki epoch, whereas W. H. Smyth et al. (
2011
) is based on the Voyager 1 Jupiter flyby in 1979, ground-based observations from 1991, and the Galileo J0 observation in 1995. A. Moirano et al. (
2025
) suggested that the IPT observed by Voyager 1 in 1979 had a lower peak electron density in the warm torus by ∼1000 cm
−3
compared to the average density in 2016–2022. Results from S. Schlegel & J. Saur (
2023
) and A. Moirano et al. (
2025
) suggest that the IPT in 2005–2007 was potentially less dense than during the Juno epoch. Even though these do not directly suggest that the pointing was different between the Voyager epochs and the post-Cassini epochs, it is still worth noting that the IPT morphology seemingly transitions between multiple states over time.
5.2. Reconstructing the Equatorial 2D Morphology of the IPT
The best fit (
fit
) represents a typical LT dependence of the Io-FTMC. Since
fit
covers a full 24 hr LT range,
fit
can be a proxy for the distortion of the IPT due to the dawn-to-dusk electric field. By comparing
fit
and an IPT density model, we reconstruct the position of the FTMC ribbon on the equatorial plane to determine the amount of the IPT shift. We use an IPT density model, which is based on the diffusive equilibrium (E. G. Nerney
2025
; hereafter NER25 model). The NER25 model requires input equatorial radial density and temperature profiles of charged particles. In this study, we use the radial model profile from Figure 1 in E. G. Nerney et al. (
2025
): this profile is an empirical profile of the dusk region from the Voyager 1 Plasma Science inbound observations for the cold torus and ribbon and the Cassini Ultraviolet Imaging Spectrograph (UVIS) observations for the warm torus. Figure
(a) shows the empirical mass density distribution on the centrifugal equator for each ion species as well as the total value. We then obtain a steady state diffusion equilibrium of the charged particles using the NER25 model with an isotropic Maxwellian approach. Integration of the total mass density along the magnetic field line yields a nominal FTMC profile in radial distance from 5.1 to 6.9
(Figure
(b)). The FTMC ribbon is located at a distance of 5.7
in this nominal FTMC radial profile, which corresponds mostly to the peak-density distance of S
++
and O
Figure 9.
(a) Ion mass density at the centrifugal equator taken from E. G. Nerney et al. (
2025
). (b) Io-FTMC calculated by the diffusive equilibrium NER25 model. The ion species considered in this model are indicated in the legend.
Download figure:
Standard image
High-resolution image
Next, we define 200 equally spaced LT sectors and, at each LT sector, shift the nominal FTMC profile radially, either inward or outward, such that it best matches
fit
at 5.9
. This enables us to convert the FTMC to the radial shift of the IPT from Jupiter. Figure
10
(a) illustrates this procedure with the nominal FTMC input (light blue) and the shifted profile (red), in which the radial adjustment is uniquely determined at each LT sector. The radial displacement from the nominal FTMC profile allows us to determine the FTMC ribbon location at each LT sector.
Figure 10.
(a) Schematic illustration that visualizes how we reconstruct the IPT morphology by shifting the FTMC radial profile calculated from the IPT density model. Both direction and amount of the radial shift vary with the LT sector. (b) Reconstructed FTMC ribbon locations. The initial ribbon location is at 5.7
from Jupiter (the black solid line), and the reconstructed ribbon location is indicated by the red solid line. The radial displacement at each LT sector is represented by the arrows, with its length scaled by a factor of 3 for clarity.
Download figure:
Standard image
High-resolution image
Interestingly, the radial displacement (Figures
10
(b) and
11
(a)) is much more notable in the afternoon sector, where the IPT is expected to shift inward in the magnetosphere. The reconstructed IPT morphology is a result predominantly of inward shifts in the predusk sector (12:00-18:00 LT) rather than of a parallel shift of a concentric circle toward the predawn sector. Figure
11
(b) shows the ribbon locations projected on the
JSO
axis, simulating the geometry of the ground-based observations. The reconstructed ribbon locations are at
JSO
= −5.76 ± 0.29
and 5.21 ± 0.34
from Jupiter in the dawn and dusk sectors, respectively. The uncertainty is derived directly from the 3
fit errors of
fit
. Figures
11
(c) and (d) shows the best-fit ribbon distance measured from Jupiter as a function of the ansa System III longitude, where our results are compared with the ground-based observations of the S
emission ribbon by N. M. Schneider & J. T. Trauger (
1995
; denoted as “ST95”), W. H. Smyth et al. (
2011
; “SM11”), C. Schmidt et al. (
2018
; “SC18”), and H. Kondo et al. (
2024
; “KO24”). To measure the distance of the S
emission ribbon from Jupiter, N. M. Schneider & J. T. Trauger (
1995
) and W. H. Smyth et al. (
2011
) used optical 673.1 nm S
emissions, and the other two (C. Schmidt et al.
2018
; H. Kondo et al.
2024
) incorporated 671.6 nm S
emissions as well. Ansa System III longitude represents the longitude of the line of sight. The ribbon distance changes with the System III longitude because of the offset and tilt of the magnetic moment axis with respect to Jupiter’s rotation axis. On the dawn side, the reconstructed ribbon location agrees with all four ground-based observations within the 3
confidence level. On the dusk side, all four previous studies reported an inward ribbon distance compared to the dawn side, but our result suggests a ribbon closer to Jupiter, by ∼0.4
when the best fits are compared. The upper limit still agrees with some of the ground-based observations at some longitudes, but generally, our results indicate a much farther inward ribbon in the dusk ansa.
Figure 11.
Reconstructed ribbon position with the 3
uncertainty from
fit
(the red solid line with pink shade) in the JSO coordinate. The reconstructed ribbon position is compared with an initial ribbon location at 5.7
from Jupiter (the black solid line) in panels (a) and (b), which illustrate a view from the top and the Sun, respectively. The reconstructed ribbon position measured from Jupiter is also compared with the four ground-based observations in panels (c) and (d).
Download figure:
Standard image
High-resolution image
The determined ribbon position highly depends on the model assumptions, especially on the local gradient of the FTMC radial profile near Io’s orbit. A steeper FTMC profile would reduce the expected shift, and thus the expected dusk ansa ribbon position would be closer to the directly observed position. Since the FTMC is approximately proportional to
, we carefully evaluated the uncertainty of the local gradient of
as a proxy for the uncertainty of the local gradient of the FTMC. We used the values from E. G. Nerney et al. (
2025
), who derived the ion density radial profile from the ground-based observations of the IPT. The details of this uncertainty assessment are described in the appendix section. The known uncertainties of the ion density local gradient add an error of only ±0.088
to the determined ribbon position, which is not sufficient to compensate for the ∼0.4
discrepancy found on the dusk. A point to note is that the uncertainties of the ion temperature radial profile are not available for this sensitivity analysis, which quite likely results in an underestimation of the FTMC uncertainty.
LT variability of the radial FTMC profile cannot be inferred in this study and should be investigated with spectroscopy of the entire plasma torus and disk. Figure 2 in C. Schmidt et al. (
2018
) shows observed radial brightness profiles of S
emissions at 671.6 and 673.1 nm as a function of the ansa System III longitude, which are clearly inhomogeneous between the dawn and dusk ansae. Using different FTMC profiles at each LT sector would provide more realistic results, even though the LT variation of the radial distribution of the IPT has not been fully understood, mainly due to geometric restrictions in the ground-based observations.
5.3. The Dawn-to-dusk Electric Field
In this last section, we estimate the magnitude of the dawn-to-dusk electric field based on the reconstructed IPT morphology. A simple way to model the dawn-to-dusk electric field on the equatorial plane is with a uniform field
(D. D. Barbosa & M. G. Kivelson
1983
), which can be defined by a potential of
where
is a given position vector with respect to Jupiter.
is measured counterclockwise from midnight. We replace
with (
), where
represents a given orientation of the potential gradient. The pointing of the electric field vector is expressed with
+ 12:00 LT.
The spatial extent over which the assumption of this uniform dawn-to-dusk electric field can be applied is not well understood. A statistical analysis of the Jovian synchrotron radiation showed that the dawn–dusk asymmetry of the synchrotron radiation and IPT are not significantly correlated, suggesting that the dawn-to-dusk electric field could be weaker in the radiation belt region (<2
; H. Kita et al.
2019
). However, the strength of the dawn-to-dusk electric field beyond Io’s orbit remains unconstrained.
The corotation electric field potential on the equatorial plane is expressed by
where
and
are Jupiter’s radius (1
) and the surface magnetic field magnitude (4.1 G), respectively. We assume a rigid corotation at a given distance, so
corresponds to the angular velocity of Jupiter’s rotation Ω
. The summation of the two potentials,
, represents the total electric field potential, and the equatorial equipotential contours will represent the orbits of charged particles in the IPT.
We compute the electric field potential for two cases of the pointing LT
: Case I corresponds to
= 55
5 (03:42 LT), and Case II to
= 90
5 (07:20 LT), following Smyth et al. (
2011
). Figures
12
(a) and (b) show the electric field potential on the equatorial plane (
JSO
= 0) with the magnitude of
= 10.5 mV m
−1
. Figure
12
(c) shows the radial distance of an equipotential contour line from Jupiter, which is interpreted as the orbital distance of the charged particles. The predicted orbital distance is compared with the reconstructed ribbon distance presented in Figure
11
Figure 12.
Color map of the electric field potential,
, and corresponding equipotential contours (the solid white lines) in the JSO coordinate for Case I (panel (a)) and Case II (panel (b)). The Sun is in the +
JSO
direction. The solid black lines represent the equipotential contours of
alone. The pointing LT of
and its perpendicular direction (±6 hr) are indicated with the dashed lines. All contour lines intersect the radial distances of 2.0, 4.0, 6.0, 8.0, and 10.0
at 6 hr later from the pointing LT. Panel (c) shows the orbital distance of the IPT charged particles as a function of Io’s LT. Equipotential distance is compared with the reconstructed ribbon position from Figure
11
Download figure:
Standard image
High-resolution image
Comparing Figures
12
(a) and (b), different field orientation results in the distortion of the equipotential contours toward different LTs. Case I, in which the uniform dawn-to-dusk electric field pointing toward 15:42 LT, best matches the reconstructed ribbon position. In contrast, if the dawn-to-dusk electric field is oriented toward 19:20 LT (Case II), the predicted orbital distance deviates from the reconstructed ribbon distance, particularly in the late-morning sector (the blue dashed–dotted line in Figure
12
(c)). We conclude that the predusk orientation of the electric field explains the LT dependence of the Io-FTMC revealed by the footprint observations in the Juno era. Our results do not rule out the postdusk orientation found in the Voyager and Galileo eras (W. H. Smyth et al.
2011
) because, as discussed in Section
5.1
, the IPT may transition between multiple steady states over time. Future studies should investigate whether the direction of the dawn-to-dusk electric field changes over time and, if so, identify the physical processes responsible for such changes.
The magnitude of the uniform dawn-to-dusk electric field is also evaluated; the best fit is
= 10.5 mV m
−1
, and by taking the
fit
uncertainty into account,
is in a range of 3.0–20.0 mV m
−1
. Since
is the field magnitude toward 15:42 LT, the projection onto the
JSO
axis,
0,YJSO
, can be compared directly with the values from the previous ribbon observations; the projected best-fit value is
0,YJSO
= 8.65 mV m
−1
and in a range of 2.47–17.3 mV m
−1
. Using the ribbon brightness ratio between the dawn and dusk ansae observed by Hisaki, the average electric field magnitude was estimated to be 3.8 mV m
−1
(G. Murakami et al.
2016
) and 2.8 ± 1.2 mV m
−1
(H. Kondo et al.
2024
). In addition, H. Kondo et al. (
2024
) estimated the magnitude of 3.9 ± 0.8 mV m
−1
using the 671.6 and 673.1 nm S
emission observed by the Tohoku 60 cm telescope (T60). A model estimate by (Y. Nakamura et al.
2023
) suggested the electric field magnitude of 15 mV m
−1
at dawn and 12 mV m
−1
at dusk can be generated through the magnetosphere–ionosphere coupling, with a particular enhancement of the ionospheric conductance due to meteoroid influxes. This modeling work addressed the discussion on the formation process of the dawn-to-dusk electric field. However, the relatively large uncertainties in our result do not yet allow us to place additional constraints on the topic.
Since the determination of
0,YJSO
is based on the reconstructed FTMC ribbon positions, the low precision of the determined
0,YJSO
reflects the limitation of our method using the footprint lead angle. The uncertainty of
0,YJSO
is predominantly due to the fitting error of
fit
, whereas the local gradient of the FTMC radial profile adds a minor contribution to the overall error as described in the previous section. In spite of the errors of the derived FTMC that propagate from the initial errors in lead-angle measurement, the derived FTMC exhibits large deviations (Figure
). The first-order Fourier fit (
fit
) reasonably captures the LT dependence of the FTMC and the orientation of the dawn-to-dusk electric field, but one should be careful because
fit
and its errors are also influenced by temporal variation. The magnitude of the dawn-to-dusk electric field varies in response to the temporal changes in the solar wind dynamic pressure: G. Murakami et al. (
2016
) reported that the dawn-to-dusk electric field increased to 8.6 mV m
−1
when the solar wind dynamic pressure was enhanced. This suggests that variation in the solar wind dynamic pressure can increase the dawn-to-dusk electric field by a factor of ∼2.
Temporal variation in the IPT ion density and fraction should also be significant. R. Hikida et al. (
2020
) investigated a volcanic event in 2015 using Hisaki’s spectroscopic observations of the IPT. The total ion mass density can be estimated by Σ
, where
is the electron number density, and
and
represent atomic mass and mixing ratio of ion species
(S
, S
++
, S
+++
, O
, O
++
, and H
), respectively. According to the electron density and the ion mixing ratios derived by R. Hikida et al. (
2020
), the total ion mass density increased by ∼43% in the dawn ansa and by ∼50% in the dusk ansa due to Io’s volcanic activity (Table
). The Io-FTMC we retrieved, on the other hand, exhibits approximately a 50% increase from
fit
at a given LT (Figure
), which is comparable to the expected increase of the total mass density due to eruption. Even though the FTMC depends on both ion mass density and temperature (Equation (
17
)), this comparison suggests that the FTMC can be used as a proxy for the temporal variability of the IPT. Future work should analyze correlations between the Io-FTMC and Io’s volcanic activity during the Juno era. By subtracting transient variations due to Io’s eruptions, we expect to further minimize the uncertainties in the estimation of the magnitude of the dawn-to-dusk electric field.
Table 2.
Total Ion Mass Density in the 2015 Volcanic Event Computed with the Parameters of R. Hikida et al. (
2020
2015 Jan 3 (DOY 3)
2015 Feb 7 (DOY 38)
2015 Apr 1 (DOY 91)
State
Volcanically quiet
Core electron density at maximum
Hot electron fraction at maximum
Total ion mass density [10
AMU cm
−3
2.14 ± 0.346 (Dawn) 3.06 ± 0.626 (Dusk)
3.07 ± 0.500 (Dawn) 4.56 ± 0.855 (Dusk)
2.34 ± 0.425 (Dawn) 2.60 ± 0.291 (Dusk)
Download table as:
ASCII
Typeset image
6. Conclusion
The Juno-UVS observations have provided an extensive dataset of UV images of Io’s auroral footprints in Jupiter’s atmosphere. Unlike the previous studies that used the ground-based observations, we analyzed the lead angle of Io’s auroral footprints to investigate the dawnward shift of the IPT and the dawn-to-dusk electric field in the Jovian magnetosphere.
A total mass column density in the Io flux tube (Io-FTMC) was estimated from the equatorial lead angle of the Io footprints that were observed by Juno-UVS. It is revealed that the Io-FTMC is correlated with Io’s LT in the Jovian magnetosphere, presumably due to the dawn-to-dusk electric field. Our detailed analysis of the LT dependence of the Io-FTMC allowed the derivation of the orientation of the dawn-to-dusk electric field without the limitation of an Earth-based viewing geometry. The magnitude of the dawn-to-dusk electric field was also determined, with the upper limit exceeding previous estimates.
The main conclusions of this study are listed as follows.
1.
New empirical formulae for the equatorial lead angle of Io’s MAW and TEB spots were derived from Juno-UVS data by taking the physical size of the MAW spot into account.
2.
The ion number density and temperature at the orbit of Io were retrieved from the measured equatorial lead angle by computing the Alfvén wave travel time in the MAW, as well as the TEB travel time between the hemispheres. The retrieved parameters are degenerate, but some results with weak degeneracy yield the best-fit density and temperature that are within the ranges reported by M. G. Kivelson et al. (
2004
).
3.
A field-line integrated mass column density, the FTMC, was calculated with the retrieved ion number density and temperature. With the careful sensitivity analysis, it was revealed that the Io-FTMC is a reasonable proxy for the variation of the plasma condition around Io. The mean Io-FTMC over the Juno epoch is 1.50 × 10
−10
kg m
−2
= 5.71 × 10
−3
kg Wb
−1
, which is comparable to ∼2.5 × 10
−3
kg Wb
−1
in the Galileo epoch (F. Bagenal et al.
2016
).
4.
A clear correlation between the Io-FTMC and Io’s LT has been revealed. This LT dependence is a result of the radial shift of the IPT due to the dawn-to-dusk electric field. The peak FTMC is found when Io is located at 03:42±00:53 LT, which indicates that the dawn-to-dusk electric field is oriented toward the predusk sector. This contrasts with the previous works that indicated the near-true (C. Schmidt et al.
2018
) and postdusk (W. H. Smyth et al.
2011
) orientations of the dawn-to-dusk electric field. The predusk orientation accounts for the LT dependence of the Io-FTMC during the Juno era, whereas the postdusk orientation found in the Voyager and Galileo eras (W. H. Smyth et al.
2011
) could suggest the temporal variation of the orientation.
5.
The ribbon position was reconstructed from the derived Io-FTMC by shifting a nominal FTMC radial profile, which is based on the equatorial profile of density and temperature from Voyager 1 and Cassini UVIS observations. The field-aligned density distribution is calculated using the Io torus density model (E. G. Nerney et al.
2025
). In the dawn ansa, the reconstructed ribbon position agrees with the ground-based observations within the 3
confidence level. In the dusk ansa, however, the reconstructed ribbon is located further inward by approximately 0.4
, which cannot be fully explained by known uncertainties in the FTMC radial profile. This discrepancy highlights the limitations of inferring ribbon positions solely from footprint lead angles and suggests caution in interpreting the absolute ribbon positions derived here.
6.
The magnitude of the dawn-to-dusk electric field was also estimated at 3.0–20.0 mV m
−1
toward 15:42 LT. The magnitude is projected onto the true dawn–dusk, 2.47–17.3 mV m
−1
, whose upper limit still exceeds the previously observed values. The large uncertainty of the determined magnitude may reflect temporal variations in the solar wind dynamic pressure. Temporal changes in the plasma conditions driven by Io’s volcanic activity are also likely to contribute. A future study is expected to further reduce the uncertainties in the estimation of the magnitude of the dawn-to-dusk electric field by accounting for the transient effects of Io’s eruptions.
As proven by C. Schmidt et al. (
2018
), it is still quite challenging to evaluate the orientation and spatial extent of the dawn-to-dusk electric field through ground-based observations because of the geometric restriction. Even with space telescopes, such as the Hubble Space Telescope and Hisaki, the same geometric restriction prevents one from analyzing the electric field in the noon–midnight direction. JUICE and Europa Clipper, both now en route to the Jovian system, will orbit close to the equator, providing good opportunities to map the direction and speed of the azimuthal plasma flow inside the plasma torus and disk, thanks to their orbits with low inclination with the equatorial plane. JUICE/RPWI will also measure the DC electric field, which we expect to add valuable information to understand the spatiotemporal structure of the dawn-to-dusk electric field.
Acknowledgments
S. Satoh thanks E. G. Nerney, J. Rabia, C. K. Louis, and K. Saito for fruitful discussions. S. Satoh acknowledges support from JSPS KAKENHI grant No. 24KJ0433 and 22K21344 and from MEXT WISE Program for Sustainability in the Dynamic Earth. S. Satoh and V. Hue acknowledge the support of CNES to the Juno mission. V. Hue acknowledges support from the French government under the France 2030 investment plan, as part of the Initiative d’Excellence d’Aix-Marseille Université - A*MIDEX AMX-22-CPJ-04. F. Tsuchiya acknowledges the support from JSPS KAKENHI grant No. 20KK0074. A. Moirano was supported by the Fonds de la Recherche Scientifique—FNRS under grant(s) No. T003524F.
Appendix A: Sensitivity Analysis of the Plasma Parameter Retrieval Method
As described in Section
3.3
and Table
, both the ion density (
) and temperature (
) spaces sufficiently cover the ranges based on the in situ observations (M. G. Kivelson et al.
2004
). When strong degeneracy between
and
is present in the retrieval result, the 3
confidence level extends across almost the entire range of the
and
axes and does not close within the chosen parameter space. In such cases, the chi-square minimum
is found at the upper boundary of either
or
. For instance, in the PJ3 subset,
is located at the upper boundary of
; (
) = (5000 cm
−3
, 44.0 eV) (Figure
A1
(a), hereafter Case I). This raises a concern about whether it corresponds to the exact minimum of the grid search.
Figure A1.
(a) Heat map of the reduced chi-square as a function of
and
. The
axis is up to 5000 cm
−3
. The three contour lines represent the 1
, 2
, and 3
retrieval confidence levels. The best-fit parameter is indicated as a white dot. (b) Same as panel (a), but the
axis is extended up to 7500 cm
−3
Download figure:
Standard image
High-resolution image
To address this, we extend the
axis up to
= 7500 cm
−3
to evaluate the influence of increasing the upper
boundary on the retrieved best-fit parameters (Figure
A1
(b), Case II). Figure
A1
(b) shows the heat map of the reduced chi-square with the extended
axis; the new minimum is located at (
) = (5395 cm
−3
, 40.7 eV).
can be converted into the electron density
by
where
= 1.4 is the ion average charge. The upper boundary
= 5000 cm
−3
(Case I) is therefore equivalent to
= 7000 cm
−3
, and the exact best-fit
= 5395 cm
−3
(Case II) to
= 7553 cm
−3
. Both
= 7000 cm
−3
or 7553 cm
−3
are notably high compared to the ranges from the previous studies. M. G. Kivelson et al. (
2004
) reported a range of
= 1200–3800 cm
−3
, and K. Yoshioka et al. (
2018
) and R. Hikida et al. (
2020
) reported up to ∼2900 cm
−3
when Io was volcanically active in 2015. However, with such strong degeneracy, we cannot obtain a uniquely determined solution because the parameters within the 3
confidence level show no statistically significant difference. This shows the limitation of our method to derive the ion density and temperature solely from the footprint lead angle.
Nonetheless, the FTMC overcomes this limitation and is a good proxy for the variation of the IPT. Figure
A2
(c) compares the histogram of the Io-FTMC obtained from the two cases. The light-pink bins represent the contribution that is added by the extension of the
axis. The red bins, on the other hand, are shared by both cases. Figure
A2
(c) shows that extending the
axis increases the uncertainty of the FTMC calculation. However, the median Io-FTMC, obtained as described in Section
4.2
, changes only by 0.2%: the median FTMC is 1.556 × 10
−8
kg m
−2
and 1.559 × 10
−8
kg m
−2
for Cases I and II, respectively. Additionally, the minimum in Case I remains within the 3
confidence level from Case II. We therefore conclude that there is no statistically significant difference between the two minima and that the best-fit values found at the upper boundary are not an issue in the derivation of the Io-FTMC.
Figure A2.
(a), (b) Heat maps of the FTMC as a function of
and
. The
axis is up to 5000 cm
−3
and 7500 cm
−3
in panels (a) and (b), respectively. The best-fit parameter is indicated as a white dot. (c) Histogram of the Io-FTMC obtained from Case I (red bins) and Case II (red and light-pink bins).
Download figure:
Standard image
High-resolution image
Lastly, we determine how the two constants in the lead-angle fit, the charge
and electron density
, would affect the retrieval results. While these quantities vary, we get the same Io-FTMC from the lead-angle fit; the scale height
(Equation (
)) derived in the lead-angle fit will not change, nor will the Io-FTMC (Equation (
16
)), even if
and
vary, because changes in
and
are compensated by
. Moreover, the change of
due to the variability of
and
is also quite smaller than the retrieval uncertainty because the product
is not a dominant term for
: if
is reduced from 6 to 4 eV, the best-fit
will increase from
eV to
eV for the PJ19 subset and from
eV to
eV for the PJ28S subset. Therefore, we believe that assuming the typical values for
and
is reasonable in this work.
Appendix B: Sensitivity Analysis of the Dawn–Dusk Asymmetry in the IPT Radial Profile
The determination of the ribbon shift on the dawn and dusk sides highly depends on the local gradient of the FTMC radial profile. The fact that our study suggests a large discrepancy between the previously observed ribbon positions on the dusk side motivated taking a careful assessment of the uncertainties of the FTMC radial profile.
The aim of this section is to derive the sources of uncertainties in the FTMC radial profile in order to interpret this study’s results on the reconstructed ribbon positions, despite having limited observations of such quantities. As described in the main text, the FTMC is approximately proportional to
, and thus the uncertainties of the nominal FTMC profile (shown in Figure
(b)) can be assessed by the uncertainty of each parameter
and
. However, the uncertainties of the ion temperature parallel to the magnetic field line, thus directly related to
, are poorly constrained. Here, we only determine the uncertainties of
based on the ground-based observations of the IPT, while noting that this would ultimately underestimate the FTMC uncertainty.
Here, we have the nominal FTMC radial profile
as a function of the radial distance
, and the local gradient of
) is defined by
where
and
(Figure
B1
). Since the same nominal FTMC radial profile (
)) is radially shifted at each LT sector to determine the ribbon position,
and
can be replaced with
and
Dusk
Dawn
, respectively. The radial shift on the dawn side is only 0.06
, which enables approximating
Dawn
∼ 5.9
, then
represents the radial shift on the dusk side. We therefore assess the displacement of
to evaluate the radial shift on the dusk side:
Figure B1.
Definition of
and
. The local gradient of the FTMC is
Download figure:
Standard image
High-resolution image
Its fractional change is represented by the fractional change of
Assuming that the ion temperature
does not depend on the LT, Equation (
B1
) can be replaced by:
where
represents the local gradient of the ion mass density
, and
is defined by
. Using Equations (
B3
) and (
B4
), the fractional change of
can be determined by the fractional change of
E. G. Nerney et al. (
2025
) derived the radial density profiles of the S
, S
++
, and O
ions from the ground-based observations of the IPT. They used the diffusive equilibrium model to infer the field-aligned density distribution and conducted the plasma diagnostics using the CHIANTI atomic emission database. The uncertainty in the dawn–dusk asymmetry was therefore assessed using the above error propagation, and by calculating the ion mass density (i.e., S
, S
++
, and O
densities) using Figure 14 in E. G. Nerney et al. (
2025
). Figure
B2
shows the total ion mass density around Io’s orbital distance at 5.9
. To assess the local gradient of the ion mass density, a linear fit was performed on this radial profile, specifically between 5.78 and 6.14
as the distance of interest. We then obtained the following line-fit;
Figure B2.
Ion mass density as a function of the distance from Jupiter
. The observed values were calculated from Figure 14 in E. G. Nerney et al. (
2025
). The best line fit is shown with the black solid line (see Equation (
B6
)). The vertical dashed line shows the orbital distance of Io at 5.9
Download figure:
Standard image
High-resolution image
The fractional change of the local gradient, Δ
∼ 0.18, directly leads to the same fractional change for
(Equation (
B5
)). Hence, the uncertainty of
leads to Δ
∼ 0.088
on the dusk side, where the shift in the reconstructed ribbon is
∼ 0.49
(Figure
11
). Δ
∼ 0.088
is only about one quarter of the error that originates from
fit
(±0.34
) on the dusk side. Since the assumption that the ion temperature
does not depend on the LT was made, the uncertainty of Δ
∼ 0.088
is quite likely underestimated.
Unfortunately, neither UV nor visible spectroscopy of the IPT allows deriving the ion temperature since the volume emission rate of the ion emissions does not depend on the ion temperature (but on the electron temperature). Imaging observations of the IPT have enabled one to derive the ion temperature parallel to the magnetic field line from the scale height of emission line for S
(e.g., N. M. Schneider & J. T. Trauger
1995
). However, no existing observations enable the estimation of the emission scale height for all major constituent ion species at this moment. Future continuous in situ measurements of the ions would help understand the spatiotemporal variations of the ion temperature in the IPT.
Please wait… references are loading.
10.3847/PSJ/ae3678