Search for Jovian decametric emission induced by Europa on the extensive Nançay Decameter Array catalog

Context. The electrodynamic interaction between the Galilean satellites and the Jovian magnetosphere generates Alfvén wings that connect the satellites to the polar atmosphere of Jupiter and induce auroral radiation through the cyclotron-maser instability. The satellite control of the Jovian decametric emission is widely known and has been studied since the 1960s, being ﬁrst discovered with regard to Io and, more recently, Ganymede. The partial control of these emission by Europa and Callisto, however, has not yet been conﬁrmed, however, hints of this control have already been found. Aims. The goal of this work is to search for evidence of control of the Jovian decametric emission by the satellite Europa. Methods. For this purpose, we analyzed the extensive digital catalog of Jovian decametric emission detected by the Nançay Decameter Array from 1990 to 2020. We analyzed distributions of the occurrence probability of the emission not induced by Io nor by Ganymede as a function of Europa phase and of the Array’s longitude with regard to the Jovian central meridian of longitude. Results. As a result, we selected 267 possible Europa-induced emission, from which 186 are from source A (Eu-A), 56 are from source C (Eu-C), and 25 are from source D (Eu-D). The general maximum frequency and duration of these emission are presented and compared to those of the other emission in the catalog and their average power is estimated as a function of the average power of the Io-induced emission. Conclusions. We conclude that Europa, just as in the case of Io and Ganymede, induces a portion of the Jovian decametric emission.


Introduction
Jovian decametric (DAM) radio emission are the only known type of planetary radio emission possibly observed by both space-based instruments and ground-based observatories due to the powerful magnetic field of Jupiter, which enables the generation of emission with maximum frequency up to 40 MHz, via the cyclotron-maser instability (CMI; Zarka 1998).This fact enabled the discovery of Jupiter as a radio source (Burke & Franklin 1955) even before the start of spacecraft launches for in situ exploration.Further, it led to the inference of the Jovian strong magnetic field and large magnetosphere, and to the development of early studies regarding the nature of the emission and the origin mechanism.Soon after the discovery of Jupiter's radio emission, the partial dependence between the emission occurrence and the orbital phase of the satellite Io was identified (Bigg 1964) and enforced the idea of an electromagnetic connection between Io and Jupiter that would grant the control of part of the emission by the satellite (Piddington & Drake 1968;Goldreich & Lynden-Bell 1969;Goertz 1980;Neubauer 1980).
Given that the Galilean satellites other than Io (e.g., Europa, Ganymede, and Callisto) present some properties in common with Io, such as their entire orbits being situated within the

Instrumentation and data
The NDA is a phased array located in the Station of Radioastronomy of Nançay, France, built in the 1970s for the study of decametric radio emission from the Sun and Jupiter, becoming operational in late 1977.These two targets are observed by the NDA daily for up to 8 h, approximately 4 h around their meridian transit.The NDA consists of 144 helicoidal antennas sensitive to the frequency range of 10 MHz-120 MHz, making it possible to carry out a measurement of the greater part of Jovian DAM emission.The antennas are distributed over an area of 8000 m 2 , which assures the high sensitivity of the array, in two subarrays of 72 antennas each, with opposite senses of circular polarization, which enables us to distinguish the dominant polarization of the emission (Boischot et al. 1980;Lecacheux 2000;Lamy et al. 2017).This array was built to be contemporary to the Voyagers Planetary Mission in order to concomitantly study Jupiter with the spacecraft.
The NDA is simultaneously connected to more than one receiver in order to obtain observations in different time and spectral resolutions.The receivers are spectral analysers that have been digital since 1990, namely: the Routine, the New Routine, the Mefisto, and the JunoN.Routine is the oldest receiver still connected, since 1990, and measures the flux density of the emission with dominant circular or elliptical polarization in the left-hand (LH) sense or the right-hand (RH) sense at time-frequency resolutions of 500 ms × 75 kHz for Jupiter, and 500 ms × 175 kHz for the Sun (Lamy et al. 2017).
The extensive NDA catalog of Jovian DAM emission (currently, ∼31-yr long, from 1990 to 2020) comprises the daily observations of Jupiter through the Routine receiver.The emission is visually detected on the dynamic spectra and is collected by manual selection through polygonal demarcation.The catalog presents the emission with the corresponding ephemeris data, namely: the longitude of the Earth-Jupiter line, the phase, and the longitude of the Galilean satellites and of Amalthea, as well as the Sun's and the Earth's jovicentric latitudes (Marques et al. 2017), and the measured emission characteristics, such as the start and ending times of their visualization by the NDA, their duration, dominant polarization, intensity, and frequency.Each emission in the catalog is associated with one value of intensity, which corresponds to the average, calculated after removing the background intensity; and with several values of frequency, which correspond to the maximum and minimum frequencies of the emission through time.In a ∼8 h-long observation, we have none, one, or more than one emission that may be detected.The catalog was first assembled by Marques et al. (2017) and has been constantly updated since then.

Search for possible Europa-induced emission in the NDA catalog
In this work, the search for the possible control of part of the Jovian DAM emission by the satellite Europa is performed through the analysis of occurrence probability of the emission on the NDA's digital catalog, as an extension of the work of (Zarka et al. 2017(Zarka et al. , 2018)).The control of the emission by a Jovian satellite is indicated by non-uniform distributions of occurrence of the emission along the satellite's orbital phase angle (Φ Sat ) and along the observer's longitude relative to the Jovian Central Meridian Longitude (CML).The Φ sat is counted counterclockwise from opposition to the observer (for more on the angles definition, see Fig. 1 of Marques et al. 2017).
Due to the beaming geometry of the Jovian DAM emission, via the CMI beaming hollow cones along the Jovian magnetic field lines (Wu & Lee 1979;Dulk 1985), the emission become visible to an observer located at the Earth only when the beaming cone's axis points to the dawn or dusk limbs of Jupiter with regard to the Earth's location, configuring the classic definition of DAM sources: A, B, C, and D (see Fig. 2 of Marques et al. 2017).The sources A and C are located in the dusk limb of Jupiter, at the northern hemisphere and at the southern hemisphere, respectively; while the sources B and D are located in the dawn limb, at the northern hemisphere and at the southern hemisphere, respectively.Moreover, the satellite control of the emission is set through field-aligned electric currents flowing between the satellite and the source region (Kivelson et al. 2004).Therefore, this control is indicated not by any non-uniform distribution of occurrence probability of the emission along the satellite phase, but by high-probability regions (or peaks) at specific phases, when the satellite is near the sources.Close to the dusk limb of Jupiter (sources A and C), the satellite's orbital phase is of ∼270 • ; while that close to the C and D sources, in the dawn limb, the satellite orbital phase is of ∼90 • , both cases considering no separation between the active field line and the instantaneous satellite field line (null lead angle δ).As a consequence, to indicate partial control of the emission by Europa, we expect to find regions or peaks of high probability of emission occurrence around Europa phases of 90 • and 270 • .
In our analyses, we excluded the emission induced by Io or Ganymede (hereafter called, respectively, the Io-DAM and Ga-DAM emission) in order to remove the effect of these satellites on the emission occurrence, which could appear as Europa control due to the orbital resonance among Io, Ganymede, and Europa.The selection of the Io-DAM emission and Ga-DAM emission on the NDA's catalog was based on the works of Marques et al. (2017) and Zarka et al. (2017Zarka et al. ( , 2018)), which defined CML × Φ Sat intervals of occurrence of these emission  In the CML × Φ Sat diagrams of Fig. 1, no phase intervals with enhanced occurrence probability of the emission are observed, which indicates that the emission is indeed not induced by Io nor by Ganymede.Instead, occurrence probability gaps are observed, such as in ∼220 These gaps result from the removal of the Io-DAM and Ga-DAM emission (Zarka et al. 2017(Zarka et al. , 2018;;Marques et al. 2017).The (CML; satellite phase) regions associated with each component of the Io-DAM emission and the Ga-DAM emission are shown in Fig. 1 of Zarka et al. (2018), in which a pattern can be noticed in both CML range and satellite phase range associated with each component of the satellite-induced emission.The satellite-induced emission from Jupiter's dusk limb (source A or C) occur normally at high values of longitude, from ∼180 • to ∼360 • for source A, and from ∼260 • to ∼460 • for the source C, both at Φ Sat between ∼180 • and ∼260 • .On the other hand, the satellite-induced emission from the dawn limb (source B or D) occur in Φ Sat < 130 • , between the longitudes of ∼60 • and ∼200 • , for the source B, and in a wide longitude range for the source D.
The number of events for each component of the Io-DAM, the Ga-DAM, and the non-Io-Ga DAM emission in the current NDA's catalog, as well as the median, average, standard deviation, and maximum values of their maximum frequency and duration, are shown in Table 1.As observed by Marques et al. (2017), the Io-DAM emission have in general higher maximum frequencies than those of the non-Io emission, mainly for the components A and B. On the other hand, the Ga-DAM emission have, in general, maximum frequency similar to the ones of the non-Io-Ga emission.With regard to the emission duration, all components of the Io-DAM emission seem to be longer than the Ga-DAM and non-Io-Ga emission, which, in turn, have about the same duration within the statistical uncertainty.The Ga-DAM emission from the source A (Ga-A) are in average 6 min shorter than the non-Io-Ga A emission, and the Ga-C emission, >10 min longer than the non-Io-Ga C emission.
Moreover, the occurrence probability of the non-Io-Ga emission in the CML × Φ Sat was calculated with regard to Europa's orbital phase to search for indications of control of the emission by Europa.The plots are shown in Fig. 2. The emission were also separated by their sources at Jupiter, and their occurrence probability was determined for every 5 • × 5 • bin, as described before.Additionally, for each component of the non-Io-Ga DAM emission, we also obtained the distribution of occurrence probability integrated over the entire CML interval (plots in the center of Fig. 2) and over the CML range with highest probability (on the right).This configuration is applied for better evaluating the peaks' amplitude along the satellite phase, with the occurrence probability presented as a function of the standard deviation (σ) of the set.The CML ranges of integration are indicated on top of the plots and by the dashed lines in the (CML, phase) diagrams.From these plots in Fig. 2, we selected high-probability regions (CML × Φ Eu ) that may be associated with Europa-induced (Eu-DAM) emission occurrence.The intervals of Europa phase associated with these regions are colored in the plots.
In both the data sets relative to the source A (Fig. 2, plots a and b) and to the source C (plots g and h), distributed and integrated over the entire CML range, two high-probability regions were observed, revealing a quasi-harmonic behavior of the emission occurrence probability when it is plotted as a function of Europa phase.For source A, the highest peaks are observed around Φ Eu = 50 • and Φ Eu = 250 • .For source C, the highest peaks are observed around Φ Eu = 30 • and Φ Eu = 250 • .This quasi-harmonic behavior is not expected, since sources A and C are both located at the dusk limb of Jupiter; close to this location, the satellite is at an orbital phase of approximately 270 • (for a null lead angle, δ = 0 • ).Therefore, although the highprobability peaks observed within 0 • < Φ Eu < 60 • in plots b and h seem to indicate control of the emission by Europa, they are not related to the occurrence of Eu-DAM emission, since Europa is close to the opposite limb of Jupiter, while the emission are beamed from the source A or C. On the contrary, we found that this quasi-harmonic distribution results from the removal of the Io-induced emission from the sources A and C combined with the Laplace resonance between the satellites, as shown in the analysis presented in the Appendix A.
Furthermore, when we integrate the probabilities over the CML ranges of 200 • -290 • for the source A (plot c) and of Relative to source B (Fig. 2, plots d, e, and f), no relevant high-probability peak was observed, which does not exclude possible Europa control of the emission from the source B; however, this does indicate that if Europa indeed controls part of these emission, there are still not enough samples of the Europa-induced emission to stand out among the non-Io-Ga B emission.
Finally, for source D, one thin high-probability peak was observed when the emission occurrence probability was A67, page 5 of 13 A&A 665, A67 (2022) integrated over the entire CML range (Fig. 2, plot k), around the Europa phase of ∼210 • .However, when the probability is integrated over the CML range of 165 • -290 • , in which the occurrence probability is the highest, a well distinguished peak of probability shows up alone between 120 • ≤ Φ Eu < 150 • , as shown in plot l, when Europa is nearer to the Jovian dawn limb.This peak may be associated with the occurrence of emission from the source D induced by Europa and, thus, indicates partial control of this component by Europa.We, thus, name this peak Eu-D, as well as the phase and CML ranges associated with it.

Comparison with Europa-induced DAM emission in the catalogs of Cassini and the Voyagers
In order to test our emission selection, we also considered the results published by Louis et al. (2017), with regard to the Eu-DAM emission on the Jovian DAM emission catalogs of Voyager 1 and 2, and Cassini spacecraft.Those authors used the Exoplanetary and Planetary Radio Emissions Simulator (ExPRES) (Louis et al. 2019) to simulate Eu-DAM emission as seen by those spacecraft and then compared the simulations with the real data, classifying the matching emission as probable Eu-DAM.
The sources of the emission could also be inferred.Louis et al. (2017) analyzed emission detected by the Voyager-1 and Voyager-2 and Cassini during their flybys of Jupiter in 1979Jupiter in and 2000Jupiter in -2003, respectively, respectively.The radio experiments on both Voyager spacecraft are sensitive to the frequency range of 1.2 kHz-40.5 MHz (Warwick et al. 1977), covering the entire Jovian DAM emission frequency range.The antennas of Cassini, on the other hand, were sensitive to the frequency range of 3.5 kHz -16 MHz (Gurnett et al. 2004), covering only part of the Jovian DAM emission.Europa phase intervals relative to the Eu-DAM emission detected in their analysis were 80 • < Φ Eu < 115 • (for the emission from the sources B and D), and 245 • < Φ Eu < 280 • (for the emission from the sources A and C).These intervals are delimited in Fig. 2 by the horizontal white lines and the vertical black lines.In addition, Lamy (2016) analyzed the average flux density of Cassini's observations in its flybys of Jupiter and found a hint of Eu-D emission occurring in a phase interval coinciding with the one observed by Louis et al. (2017).
However, even though those results might contribute to endorsing part of our analysis, it is necessary to mention the differences between the data collected by the different instruments.The proximity of the Voyager and Cassini spacecraft to the sources during the flybys enables the detection of weak emission, while the NDA can detect only strong emission due to its large distance from the sources.Therefore, the spacecraft can also detect emission that is not observed in the NDA's catalog.Additionally, the frequency range observed by Cassini, whose catalog was the most extensive one analyzed in the study of Louis et al. (2017), corresponds only to the lower part of the frequency range of the Jovian DAM emission, while the NDA's observation frequency range corresponds to most of the Jovian DAM frequency range, from ∼10 MHz.Those divergences certainly affect the observed emission and, thus, the phase and CML intervals associated with the probable Eu-DAM emission.
Therefore, in the comparison with the results of Louis et al. (2017), either matching or near-matching between our selection of phase intervals and their phase selection may reinforce our selection.This comparison supports the selection of the phase intervals associated with the probability peaks Eu-A, Eu-C, and (less strongly) Eu-D, but in different CML ranges (see Fig. 6 of Louis et al. 2017).Although the Φ Eu range relative to our Eu-D emission does not coincide with that found by Louis et al. (2017), we note that the Eu-D peak is preeminent in the data set and the phase lag observed is minimal.In addition, we also note that the number of samples of non-Io-Ga DAM emission from the source D is the smallest (with max.occurrence probability of 5%) among all the components of the non-Io-Ga DAM emission.Then, we expect that, as the catalog is updated and more samples of non-Io-Ga D emission are detected, the Eu-D peak will become more well distinguished if it is indeed related to the occurrence of Eu-D emission, perhaps even lessening the phase lag.For now, we maintain the selection of the Eu-D emission, but as only possible Eu-D.

General characteristics of the Europa-induced DAM emission
Based on the preceding analyses, we selected all the emission occurring within the (CML, phase) regions Eu-  3, in the (CML, Φ Eu ) diagrams of the occurrence probability distribution of all the non-Io-Ga DAM emission and of the Eu-DAM emission alone (plots a and b, respectively).These diagrams may be compared with the ones of Fig. 1 of Zarka et al. (2018).We note that the pattern in the CML and satellite phase ranges for each component that was observed for the Io-DAM and the Ga-DAM emission is repeated for the Eu-DAM emission, mainly for the Eu-A and Eu-C components, which strengthens these emission selection.The CML and satellite phase ranges of the possible Eu-D emission coincide partially with those of the Io-D emission.The distributions of maximum frequency and duration of Europa-induced emission are shown in Fig. 4 and the statistical parameters (median, average, and standard deviation) of these distributions are shown in Table 2.The maximum frequency of the Eu-A emission (average of 24.9 MHz), originating in the northern hemisphere of Jupiter, is on average higher than the maximum frequency of the Eu-C (avg. of 19.2 MHz) and Eu-D (avg. of 20.8 MHz) emission, both originating in the southern hemisphere of Jupiter.Higher values of maximum frequency of the Jovian northern emission in comparison with those of the southern emission are expected as a consequence of the high intensity of the magnetic field in the northern hemisphere due to the presence of the anomaly of high-amplitude magnetic field (Connerney et al. 2018).
With respect to the duration of the emission, it is an aspect that is difficult to evaluate based on the NDA's catalog because the apparent duration of the emission is affected by diverse factors, such as the rotation of Jupiter and of the Earth, the change in declination of the Earth with regard to Jupiter (Boudjada & Leblanc 1992;Leblanc et al. 1993), and the distance variation between the planets (Zarka et al. 2018).These factors contribute to an accumulation of short emission.Even so, from the values of average duration on Table 2, the Eu-D type appears to be the one with the longest emission, with average of 42.0 min, followed by the Eu-C emission and the Eu-A emission, with average of 40.5 min and 33.9, respectively.We highlight, however, that A67, page 6 of 13  since the southern emission has a maximum frequency that is lower than the northern emission, the emission from sources C or D are not as well visually distinguished in the dynamic spectra as the emission from sources A or B, due to radio interference in low frequencies and the terrestrial ionosphere cut-off.This may lead to the detection of only small portions of the southern emission.Therefore, the Eu-D and Eu-C emission may be even more extended in time.
In comparison to the values on Table 1 relative to the Io-DAM and Ga-DAM and the Non-Io-Ga emission, the Eu-DAM emission have maximum frequency values similar to those of the Ga-DAM and the Non-Io-Ga DAM emission, and smaller than those of the Io-DAM emission, mainly in the components from the Jovian northern hemisphere.With regard to the duration, the Eu-DAM emission are on average similar in duration to the Ga-DAM emission and to the Non-Io-Ga DAM emission, both shorter than the Io-DAM emission.

Energetics of the Io-DAM, Ga-DAM, non-Io DAM, and Eu-DAM emission
In this section, we present an analysis of the emission energetics.This analysis is based on the one presented by Zarka et al. (2018), who compared the energy of the Ga-DAM emission in the 1990-2015 version of the NDA catalog with that of the

Fig. 1 .
Fig. 1.Occurrence probability of the emission not induced by Io nor by Ganymede as a function of the orbital phases of Io and Ganymede, and the Jovian longitude (CML) of the Earth-Jupiter line.The emission was detected by the NDA from 1990 to 2020.The probability was calculated in 5 • × 5 • bins (CML × phase) for each component of the Jovian emission (A, B, C, and D).

Fig. 2 .
Fig. 2. Occurrence probability of the emission not induced by Io nor by Ganymede, separated by their sources, distributed as a function of the CML and Europa's orbital phase (left).Distribution of the occurrence probability integrated over the entire CML range (0 • -360 • ) (center).Distribution of the occurrence probability integrated over the CML range in which enhanced occurrence is observed (right).The horizontal white lines (in the plots on the left) and the vertical black lines (in the other plots) delimit the phase intervals associated with the Eu-DAM emission detected by Louis et al. (2017).The dashed lines delimit the CML ranges of probability integration shown in the plots on the right.280 • -20 • for the source C (plot i), which are the respective CML ranges with the highest occurrence probability, that quasiharmonic behavior fades, resulting in a single high probability peak in each set, especially in the data relative to the source C. In the plots c and i of Fig. 2, these peaks are observed both around Φ Eu = 245 • .They are most probably associated with the occurrence of the emission induced by Europa in the sources A and C, and thus are referred to as Eu-A and Eu-C, respectively.

Fig. 3 .
Fig. 3. Occurrence probability of all emission not induced by Io nor by Ganymede in the NDA current catalog as a function of Europa phase and of the Jovian CML (a) and the occurrence probability of possible Europa-induced emission as a function of Europa phase and of the CML (b).The (CML, phase) regions associated with the Eu-DAM components are indicated by the white boxes.

Fig. 4 .
Fig. 4. Distributions of maximum frequency and duration of the emission classified as possible Eu-A, Eu-C and Eu-D.

Table 1 .
Statistical parameters of maximum frequency and duration distributions of the Io-DAM emission and the Ganymede-DAM emission, and of the non-Io-Ga DAM emission.

Table 2 .
Statistical parameters of the maximum frequency and duration distributions of the Europa-induced emission.