WorksPhysics and Astronomy
Analysis of a coronal hole jet in the chromosphere and the transition region with IRIS data and SDO observations
Abstract
Coronal jets are transient phenomena that occur throughout the solar atmosphere, often triggered by the activation of mini-filaments. Jets are an important field of study, as they are believed to contribute to the mass supply of the solar wind. We study a small-scale solar jet occurring in a low-latitude coronal hole, focusing on traces of the jet eruption to the lower atmosphere, utilizing chromospheric and transition region diagnostics. We also investigate the possible mechanisms and dynamic conditions leading to its onset. We analyze ultraviolet spectra from the Interface Region Imaging Spectrograph to investigate chromospheric dynamics and their connection to the transition region, using diagnostics of the Mg II k 2796.35 Å and Si IV 1393.75 Å lines, together with non-LTE inversions. We complement this analysis with multiwavelength imaging and magnetograms from the Solar Dynamics Observatory, as well as magnetic field modeling. The analysis of Mg II k and Si IV parameters reveals a coupling between chromospheric and transition region plasma dynamics. Profiles on the jet spire present deeper central reversals, larger peak separations and higher transition region Doppler velocities. The jet initiation is associated with a mini-filament-like structure, preceded by dynamic variations in the photospheric magnetic field. We find single-peaked Mg II k profiles at one of its footpoints, possibly indicating plasma drainage. The presence of a null point and a quasi-separatrix layer suggests that reconnection can play a role in the mini-filament-like structure destabilization and jet eruption. This study suggests that the signature and possibly the origin of coronal jets can be traced down to the chromosphere. The detailed analysis of the Mg II k profiles at the jet spire and the mini-filament-like structure may provide valuable observational constraints for future MHD simulations.
Introduction
Solar jets are transient, impulsive energy-release events that occur frequently throughout the solar atmosphere. Understanding the physical mechanisms that drive their formation is essential, as they are believed to contribute both to the heating of the upper atmosphere and mass supply of the solar wind [12, 24]. Observations of solar jets are commonly performed in the ultraviolet (UV), using high-resolution imaging from the Atmospheric Imaging Assembly (AIA) and line-of-sight (LOS) magnetograms from the Helioseismic Magnetic Imager, HMI [91], both onboard the Solar Dynamics Observatory (SDO) [58, 95], together with spectroscopic measurements from instruments such as the Interface Region Imaging Spectrograph (IRIS) [27].
A comprehensive understanding of these events requires combined observations that probe multiple atmospheric layers. IRIS provides several spectral lines, suited for this purpose. The Si iv 1393.755 Å and 1402.770 Å doublet lines are key emission lines for studying the dynamics of the solar transition region [42, 33]. Moreover, the optically thick Mg ii k (2796.35 Å) and h (2803.53 Å) resonance lines (wavelengths taken from CHIANTI [28]) sample plasma conditions from the photosphere to the upper chromosphere and thus provide powerful diagnostics on the lower-atmospheric dynamics [57, 54 b].
In quiet-Sun regions, Mg ii h&k profiles are typically double-peaked with a central reversal. However, there are many cases where they are single-peaked, for example in sunspot umbrae [49]; [11], where the source function is never well coupled to the local temperature, due to the low density [37]. Single-peaked profiles in plages can result from a hot and dense chromosphere or a transition region moved to a higher column mass [17]. In flares, large asymmetries and the absence of a central reversal are presumably caused by increased temperatures and electron densities [56, 48, 112] or significant unresolved upflows and downflows [87]. Fainter single-peaked profiles have also been found in filaments and specifically in low temperature regions [25]. However, interpreting these shapes in prominences has been proven non-trivial [39]. Barczynski et al. [10] found single-peaked profiles at the bases of coronal cavities in prominences, where plasma condensation or drainage toward the chromosphere is believed to take place. Detailed observations of Mg ii h&k profiles in mini-filaments are less common [65, 77]. Asymmetric profiles are generally the result of velocity gradients, related to plasma motions above the peak formation height [55] and wave propagation [35]. The effects of microturbulence and opacity can also influence the width and shape of the profiles [9, 67, 17]. However, the extent of turbulence of any scale in the chromosphere is still under study [45].
Eruptions of mini-filaments are generally believed to be one of the primary progenitors of solar jet eruptions [99]. Such structures are miniature versions of large-scale filaments and typically lie above polarity inversion lines [104]. Observations frequently reveal that components of the jet spire appear at different wavelengths and temperatures, and are not strictly co-spatial [16, 3]. These correspond to cool and hot plasma flows guided along different magnetic field lines but linked to the same eruptive event [20].
Magnetic reconnection is believed to be the fundamental driver of solar jets (Shibata et al. 1992; Archontis et al. 2010; Pontin & Priest 2022). Spectroscopic evidence of reconnection can be bi-directional flows, such as downflows at the Mg II k line and upflows at the Si IV line [34]. Magnetic reconnection can occur at null points, where the magnetic field vanishes, or within quasi-separatrix layers (QSLs), where magnetic connectivity changes rapidly but remains continuous [82, 100]. Studies show that jets often originate in null-point fan–spine topologies (Moreno-Insertis et al. 2008; Török et al. 2009; Kumar et al. 2019). Moreover, jets are also associated with reconnection in QSLs (Li et al. 2025), even in the absence of null points [36].
Flux emergence and flux cancellation are commonly connected to solar eruptions (Wyper et al. 2017; Török & Kliem 2005). Flux emergence has been associated with filament formation and jet activity [22, 113, 7, 21], while flux cancellation is frequently observed prior to mini-filament eruptions and coronal jets [41, 75], leading to the disappearance of both or one of the polarities at the polarity inversion line [63, 66].
This study significantly extends the work presented in Koletti et al. [50] (hereafter Paper I). Paper I focused on the transition region and coronal signatures of two small-scale jets, using Si IV and AIA diagnostics. Here, we provide a detailed spectroscopic analysis of one of the jets (jet 1) in the chromosphere using the optically thick Mg II k line. Moreover, by combining the Si IV diagnostics from Paper I with new constraints from Mg II k, we bridge chromospheric and transition region layers. This multi-height approach is essential for a comprehensive understanding of jet initiation and evolution, and highlights the central role of mini-filament dynamics in driving solar jets. Multi-height spectroscopic studies of jets and mini-filaments, particularly those including chromospheric diagnostics, remain scarce (e.g., Joshi et al. 2020; Cai et al. 2019; Panesar et al. 2025; Lu et al. 2019; Gorman et al. 2022; Madjarska et al. 2022). In this context, our work represents one of the few studies that combines high-resolution chromospheric and transition region spectroscopy to probe jet dynamics across multiple atmospheric layers, specifically focusing on the chromosphere and transition region.
The paper is organized as follows: In section 2, we present the observational data and the pre-processing techniques used in this study. In section 3 we describe the data analysis and the main results, while section 4 and section 5 discuss and summarize our main conclusions.
Observational data and analysis
Observations
The observation analyzed here took place on October 9, 2013 within a low-latitude coronal hole (CH). We used multiple IRIS datasets, including slitjaw images (SJ) at 1330 Å, 1400 Å and 2796 Å, as well as a very dense raster scan of 3.5 hours in duration and 400 slit positions. The field of view is [140.5, 181.8]″ with its center located at [X, Y] = [435.2, 302.6]″ from the center of the solar disk, and the observation covers the time interval from 16:56 UT to 20:26 UT. We used level-2 data, which have been corrected for instrumental effects such as flat-field variations and dark current, and have been wavelength calibrated using the chromospheric O I 1955.6 Å line. IRIS provides very high spatial (0.166″-0.35″) and spectral (40 mÅ) resolution data in the far-ultraviolet (FUV 1332 Å – 1407 Å).
Moreover, we used EUV data and LOS magnetograms, as recorded by AIA and HMI, respectively. The AIA data have a cadence of 12 seconds and a pixel size of 0.6″. The 304 Å channel corresponds to the emission of He II at a temperature around 7·10^4 K, depending on the dominant physical process [5]. HMI data have a pixel size of 0.5″ and cadence of 45 seconds or 720 seconds (12 minutes), with an upper bound of the random noise equal to 10 Mx cm−2 or 6.3 Mx cm−2 [62], respectively.
Data treatment
The near-ultraviolet (NUV) and FUV channels were co-aligned by comparing the fiducial lines of each detector. We also performed radiometric calibration to transform the Data Numbers (DN) into physical units (s−1 cm−2 sr−1 Å−1). All SDO images were co-aligned with IRIS with the same process as in Paper I. For visualization purposes, we took the average of every three consecutive 45s magnetograms, in order to enhance the signal. Moreover, all HMI images were enhanced with unsharp masking with a 6″ × 6″ kernel. The contrast of AIA images was improved with a mild gamma correction (γ = 1.1).
Fitting of spectral profiles
The fitting process of both the Mg II h&k and Si IV lines is based on a genetic algorithm (GA), which is an optimization technique that mimics natural selection, so that solutions evolve over time to become more optimal [40, 31]. Moreover, the profiles of both spectral lines were smoothed before fitting using the Savitzky-Golay filter, which can reduce noise while retaining the overall shape of the profile [90].
Mg II h&k
We fit the profiles with seven Gaussian curves 1: two curves for the k line, two for the h line, and three curves for the blue wing of the k line, the red wing of the h line and the middle “bump” between the k and h lines, which is formed by the red and blue wings of the k and h lines, respectively. The background is not explicitly treated, but is included in the wings and “bump”.
In this study, we focus on the Mg ii k line, with the main spectral features denoted using the standard notation: the line core is referred to as k, the violet and red emission peaks as k and k, respectively. The features are shown in Figure 1 top panel. Although only the k line is analyzed, the simultaneous fit of both lines ensures a consistent representation and precise estimation of the k-line parameters.
Figure 1. Example of a typical Mg ii h&k spectrum. Top: The GA fit and the RL improved fit. The main spectral features are indicated by arrows. Bottom: The observed profile with the individual Gaussian components overplotted.
The fitting pipeline is the following: First, we applied the least-squares (LS) method, using its parameters as initial values for the GA model, which optimizes each profile fit individually based on the highest coefficient of determination () score (Wright [107]), assigning higher weights to the k core and the k peaks to emphasize the importance of these features. The seven-Gaussian fitting helps reproduce the overall shape of the profile, distinguish the different components and reduce the effects of noise and secondary absorption lines, such as the Mn I 2795.64 Å3, which affects the blue wing of the k line. This method is also not restrictive for single-peaked profiles.
However, the algorithm is unable to simultaneously fit the k peaks and the k dip for profiles with deeper central reversals. To address this, we utilized a Reinforcement Learning (RL) algorithm, based on a Deep Q-Network (Mnih et al. 2013). The RL agent was applied to the seven-Gaussian fit, iteratively adjusting the values around the peaks and central reversals. By maximizing the weighted coefficient of determination between the fitted curve and the observed profile, the final fit was improved.
The bottom panel of Figure 1 shows an example of a fitted profile. The fitting performance is significantly improved at each stage. Using only the LS method, 45% and 39% of the profiles have and , respectively. These values improve to 8% and 47% with the GA optimization, and further to 6% and 58% with the RL adjustment. We used the RL fits in the following analysis, excluding profiles with . For double-peaked profiles, the k position is determined as the minimum between the two maxima that represent the k peaks.
Si iv 1393.75 Å
In Paper I, we presented the fitting of asymmetric profiles with a bimodal Gaussian. Here we improve this process by first fitting each line with both a single and a bimodal Gaussian curve, and then using a GA to optimize the parameters of the bimodal fit. The best model is selected based on multiple statistical criteria, including the Likelihood Ratio Test (Wilks [105]), the Akaike Information Criterion (Akaike [2]), the Bayesian Information Criterion (Schwarz [93]), and Bayes factors (Kass & Raftery [47]). If both lines strongly support a bimodal fit and the two fitted curves exhibit similarity constraints, the bimodal model is chosen; otherwise, the single Gaussian model is preferred.
Results
The section is organized into two parts. The first focuses on the evolution of a dynamic absorption feature in AIA 304 Å, probably a mini-filament, leading to the jet formation, as well as the analysis of the magnetic field and chromospheric diagnostics. The second part presents a statistical analysis of the Mg ii k and Si iv 1394 Å diagnostic parameters for the key structures discussed in Paper I.
In Paper I, we analyzed the jet evolution across multiple AIA channels. The jet first appears in the AIA 304 Å channel and subsequently in higher-temperature channels. The cool and hot counterparts of the jet spire do not spatially coincide with each other. A loop appears simultaneously, with its southern footpoint rooted at the same bright point from which the jet originates. A plume, rooted at the edges of the coronal hole, acts as foreground emission. The jet and the loop are visible in most AIA channels, as well as the Si iv spectra and 1400 Å SJI. Here, we also report a new feature: a bright redshifted structure that is clearly visible in the chromosphere but absent in higher temperatures.
Pre-eruption evolution and jet dynamics
Evolution leading to the jet eruption
Prior to the jet eruption, we observe a well-defined dark elongated absorption feature (Figure 3 arrow B). Before that, it is close to the resolution limit of the instruments, thus difficult to distinguish from its surroundings. We believe that this structure is most likely a mini-filament, as it appears in absorption in AIA 304 Å, but is invisible in the 1400 Å and 1330 Å SJ images, indicating a cool structure containing neutral hydrogen, neutral and singly-ionized helium absorbing photons of Å (Williams et al. [106]; Anzer & Heinzel [6]). Note that it is not clearly visible in AIA 193 Å and AIA 171 Å, due to the plume emission. Adams et al. [1] report on a similar absorption structure.

Figure 3. Temporal evolution of the jet as observed by multiple instruments. The green lines denote the approximate projected mini-filament-like structure tracing, as observed on AIA 304 Å. The red dashed circles at 17:23:14 and 17:24:44 show the two PILs (see text). The vertical lines on AIA 304 Å and 193 Å columns indicate the position of the IRIS slit. Arrows A: the initial position of the mini-filament-like structure (also enclosed by the white dashed circle), B: the elongated mini-filament-like structure, C: the bright redshifted feature, D: pre-existing bright features, E1: elongated feature, E2: structures possibly below the elongated feature, F1: cool jet component, F2: hot jet component, G: loop associated with jet eruption. Online animations: AIA 304 Å and AIA 193 Å, with arrows annotating selected structures.
The green lines in the HMI panels at 17:23:14 and 17:24:44 UT show the exact position of this structure, as projected in AIA 304 Å at similar timestamps. Until 17:23:43 UT, it appears as a small, dark structure roughly above a small polarity inversion line (PIL1) between a weak positive and a strong negative magnetic field patch. Its length can be roughly estimated up to approximately 3.5 Mm (see Figure 3, arrow A and online animation). It possibly forms at approximately 17:17 UT; however, it is very faint and can hardly be distinguished from the background until 17:23:43 UT, making both its exact formation time and its identification as a mini-filament uncertain. For this reason, we will refer to it as "mini-filament-like structure" hereafter.
A bright structure is observed at the same location at SJI 2796 Å (Figure 3, arrow C at 17:22:59 UT, also captured by the IRIS slit from 17:20:53 to 17:23:31 UT, evident in the Mg ii k raster at (382″, 270″), close to the southern part of the loop and the jet base (see also 3.1.5). The Mg ii profiles on this structure are almost or completely single-peaked. An example of four such profiles is shown in Figure 2. We will hereafter refer to it as "bright redshifted structure" and its exact nature will be analyzed in detail in the Discussion (section 4).
Figure 2. Four single-peaked Mg ii k profiles on the bright redshifted structure. The red line is the fitted seven-Gaussian curve and the black dashed line the actual profile (see paragraph 3.1.5).
Following again the evolution of the mini-filament-like structure, it exhibits upward motions, appearing more compact and increasing in length from 17:23:43 to 17:24:43 UT. Mini-filaments are observed to increase in length ([73]), as well as large-scale filaments ([46]). During this evolution it reaches a length of 11 Mm (15″) and its southern end is still located close to PIL1, while below its middle section a second, weaker PIL (PIL2) is also found (red circle at 17:24:44).
Simultaneously with the movement of the mini-filament-like structure, from 17:23:43 to 17:24:43 UT, we observe an elongated feature, visible in AIA 304 Å (arrow E1 on Figure 3). It forms concurrently with the mini-filament-like structure, spanning from (384″, 270″) to (395″, 285″). The high spatial resolution of the SJ 1400 Å images (arrow E2) suggests that the elongated feature possibly consists of multiple bright, fine-scale structures unresolved by AIA, or alternatively, that these structures lie below the mini-filament-like structure, thus being only partially visible.
At 17:24:55 UT, the southern end of the elongated feature (385″, 270″), also close to the southern end of the mini-filament-like structure, is intensified in AIA 304 Å (see animation). At the same time, the mini-filament-like structure continues its rising motions, until it disappears. Shortly thereafter, the cooler jet component (arrow F1) appears at 17:25:19 UT (see animation), followed approximately 1 to 2 minutes later by the hotter jet component (arrow F2) and the loop (arrow G in Figure 3), both originating from the same location ([99]).
Magnetic field evolution
The HMI 45 s magnetograms show no significant evolution of the photospheric magnetic field minutes before the jet eruption (see Paper I). To further assess temporal variations in magnetic field strength, we used HMI 720 s magnetograms up to approximately 2 hours before the jet eruption. We compute the summed magnetic field over a fixed subregion around PIL1 beneath the mini-filament-like structure ( pixels), treating positive and negative polarity separately. We also calculate the net unsigned flux, which is the sum of the positive and absolute negative flux. To avoid noise contamination, we select pixels with values greater than the nominal noise level of 6.3 Mx cm ([62]). The selected region is also kept fixed and fully encloses the magnetic polarities of interest, minimizing variability due to magnetic elements moving into or out of the box.
The top panels of Figure 4 show the temporal variation of the positive, negative and unsigned fluxes. Note that positive and negative fluxes are not balanced, because the dominant polarity of the coronal hole is negative. The bottom panels show the area within which these values are calculated with a red box in two different timestamps. The green dashed line marks PIL1.
Figure 4. Top: Temporal evolution of the integrated HMI magnetic flux with errorbars. The orange shaded area shows the duration in which we observe a rate of decrease of Mx/hour. Bottom: Three HMI 720 s magnetograms, showing the red box within which we calculate the aforementioned fluxes. The green dashed line indicates PIL1. The red arrow shows the area of decrease and shear motion of the positive polarity. An online animation of the HMI 720 s magnetograms is available.
From 15:10 UT to 15:46 UT we observe flux increase within and around the red box, which may indicate magnetic flux emergence. At 15:46 UT a positive polarity starts to move toward PIL1 and the negative polarity, leading to decrease of both polarities from 15:58 UT to 16:58 UT (orange shaded area on the top panels of Figure 4 and red arrow at 15:58 and 16:10 UT). We also observe motions of positive patches almost parallel to PIL1, possibly suggesting shear motions (red arrow 16:22 UT Figure 4). Both phenomena can also be seen in the online HMI animation. The total decrease rate of the total unsigned flux is Mx h, comparable to the observed values reported by Chen et al. [21] and simulations (Patsourakos & Archontis [79]), however slightly larger than Panesar et al. [74]. For the positive flux the rate is Mx h and the negative is Mx h. The percentage decrease is 70% for the positive and 20% for the negative flux, while for the total unsigned flux is 30%. From 16:49 to 17:11 UT, recurrent increases in intensity are observed for several minutes at the same location (see AIA 304 Å animation).
Given that the upper bound of the noise level for the 720 s magnetograms is Mx cm and assuming that all pixels are independent with each other, the total error of the summed flux is of the order of Mx. The corresponding errors for the positive and negative fluxes are of the order of Mx. All errors are smaller than the measured flux by one order of magnitude, indicating that the measured decrease is significant. However, the above values are actually lower limits, as in reality there is some level of correlation between pixels.
Magnetic field modeling
We study the magnetic topology of the observed structures by applying a potential magnetic field extrapolation to the 720 s magnetogram obtained at 17:22:16 UT. We used a method4 based on the Fast Fourier Transform approach of Alissandrakis [4]. To minimize boundary effects, the extrapolation was initially applied to a region surrounding the coronal hole that extends twice the field of view in both and directions. This way the area below the jet lies approximately at the center of the extrapolation.
Figure 5 shows the results of the above calculations. The topology provides useful insight into the magnetic connectivity of the region. We observe closed magnetic field lines of length up to 12″, spatially coincident with the loop (Figure 3 arrow G) and open lines aligned with the jet spire.
To further investigate the magnetic topology, we searched for magnetic null points, using the Null_Finder module5, which implements a trilinear interpolation and the Newton-Raphson root-finding algorithm (Haynes & Parnell [38]). Moreover, we calculated the squashing factor using the FastQSL code (Titov et al. 2002; Zhang et al. [110], which employs a GPU-accelerated adaptive Runge–Kutta–Fehlberg (RKF45) integration scheme. We found a null point at a height of 2 Mm above a photospheric parasitic (positive) polarity, also close to where the northern edge of the mini-filament-like structure is located 2.5 minutes later (Figure 5, AIA 304 Å, 17:24:43 UT). Around the null point, a small fan-like topology is evident, anchored in surrounding negative magnetic field polarities. The map of displayed in Figure 5 corresponds also to 2 Mm height. High values, (Titov et al. [100]), indicate the presence of QSLs, where reconnection is likely to occur. The null point lies near the intersection of Q structures, indicating strong changes in magnetic connectivity in its vicinity. The jet originates at the southern end of the fan surface, where regions of separate the positive from the negative magnetic polarities (red arrow in Figure 5). The apparent misalignment between the jet and the outer spine of the fan may be attributed to projection effects. Another possibility is that the potential field extrapolation may not accurately reproduce the precise location of the null point, since mini-filaments carry electric currents that can shift its position (Régnier [85]). Nonetheless, the overall magnetic topology indicates that reconnection likely plays a key role in the activation of the mini-filament-like structure and the subsequent jet eruption.

Figure 5. Results of the potential field extrapolation. The green lines correspond to open magnetic field lines, while the blue lines to closed ones. Top row: view from different angles over the extrapolated at . Bottom row, from left to right: view from elevation angle over the extrapolated at , AIA 304 Å at 17:24:43 UT and 193 Å at 17:27:42 UT. The displayed Q values are calculated at 2 Mm, with higher values shown in yellow. The yellow star denotes the null point, while green and blue lines are open and closed magnetic field lines, respectively.
Inversions of IRIS observations
To retrieve the optical-depth-dependent atmospheric properties of the jet region, we employed the IRIS inversion code (Sainz Dalda et al. [88]), based on the Stockholm Inversion Code (STiC) framework (de la Cruz Rodríguez et al. [26]). The inversion is performed with a node spacing of , using the Mg II h&k lines and the Mg II UV triplet⁶. STiC implements non-LTE radiative transfer with partial frequency redistribution (PRD) and node-based parametrization to produce a smooth stratification of temperature, velocity, density and microturbulence. These capabilities make it particularly well suited for diagnosing the dynamic chromosphere.
Inversions provide constraints on the variation of plasma parameters with optical depth (Sainz Dalda et al. [88]), beyond the profile fitting methods presented in paragraph 2.3.1. The range −6.5 < < −5 is representative of the middle chromosphere and −5 < < −4 of the low chromosphere (Sainz Dalda & De Pontieu [89]). The most reliable values of the model are those for , where is calculated at 5000 Å (Sainz Dalda et al. [88]). For this reason, we select . We also decide to include , as it is the closest optical depth where the bright redshifted structure is evident and is relatively close to −5. However, these results should be interpreted with caution. Our findings are shown in Figure 6.
Figure 6. Inversion maps of the jet region. Each row corresponds to a different value, indicated at left of each row. Arrows A: jet base, B: blueshifted stripe, and C: bright redshifted structure.
At , at the base of the jet (arrow A), we observe redshifted velocities up to 7 km/s, which could indicate plasma flows opposite to the direction of the jet eruption. Systematic redshifts are also present at the jet base at . In addition, the blueshifted strip (arrow B) possibly tracks the upward motion of the mini-filament-like structure with a velocity of 4 km/s. At the same location, we also observe lower temperatures (5000 K) and lower electron densities ( cm). At this optical depth, a few profiles exhibit redshifted velocities on the bright redshifted structure, which is also denser ( cm) and hotter (5800 K) than the surrounding regions. As already mentioned, the bright redshifted structure (arrow C) exhibits systematic redshifted velocities of 5 km/s at .
Velocities and intensities
We used the Red-Blue (RB) asymmetry to assess the plasma motions. The RB asymmetry (Leenaarts et al. [55]; Pereira et al. [80]), is the normalized difference of symmetrical integrals of
the intensity profile around the core, defined as , where and represent the integrals over the red and violet sides of the profile, respectively. Here, we set km/s as the velocity offset around the core, corresponding to the peaks, and km/s as the width of the integration window. The integrals were evaluated using the Simpson rule (Cartwright [19]).
Thus, the RB asymmetry is a measure of the variation of velocities (along the line of sight) above the peak formation height. In a simple example, suppose that the plasma of the upper chromosphere moves upward while the middle chromosphere is at rest. The photons with wavelength near will face a higher absorption probability than the ones with wavelength near . In other words, blueshifted photons emitted from deeper layers are mostly absorbed, leading to a weaker violet peak. In contrast, a downward velocity enhances the blue peak and weakens the one (see Pereira et al. [80] and Carlsson & Stein [18] for the same effect in the Ca II H&K lines).
We used the following spatial masks to identify and distinguish the different structures present in the observed region; the jet and loop masks are based on the Si IV 1394 Å Dopplergrams (see Figure 5 from Paper I), as they correspond to regions of significant blueshifted and redshifted velocities, respectively. Note that the boundaries of these regions cannot be strictly determined, and some uncertainty remains. We also note that, the "loop mask" refers to the region below the loop in Mg II diagnostics, as this structure is not visible in the Mg II k maps. The "jet mask" refers to its cooler component, since this and not the hotter component is visible in the Si IV spectra.
Figure 7 shows the logarithmic Mg II k3 intensity map, with the contours showing the jet spire, loop and background region over-plotted. The arrows point to the bright redshifted structure and the mini-filament-like structure at its ascending phase (see also arrow B on Figure 6). The RB asymmetry map and example profiles of the jet spire, loop and mini-filament-like structure are also shown. The mini-filament-like structure appears as a darker region, with lower k3 and peak intensities (see example profile). Some profiles on the bright redshifted structure are significantly brighter, as the absence of a central reversal causes the algorithm to identify the intensity maximum as k3. The RB asymmetry map shows downward motions at the bright redshifted structure in the middle chromosphere, which indicates plasma falling toward the solar surface. This result is consistent with the redshifted velocities at log derived with IRIS (see paragraph 3.1.4). At the same time, blueshifted velocities are detected at the jet spire, suggesting small upflows originating in the chromosphere.
Figure 7. Left: Map of logarithmic Mg II k3 intensity. Middle: Map of the RB asymmetry of the Mg II k line calculated over the velocity intervals and km/s. Right: Characteristic profiles corresponding to each region marked on the maps. The solid black line denotes the observed profile, while the red dashed line represents the fitted curve. Contours are color-coded as follows: blue for the jet spire, purple for the loop and black for the background region. The green and black arrows show the mini-filament-like structure (denoted with "mf") and bright redshifted structure. The blue (blueshift) and red (redshift) shaded intervals mark the spectral ranges used to compute the RB asymmetry maps. Intensities are in erg s sr cm Å.
We do not find systematic k3 Doppler velocities on the jet spire, or any indication of upward motions of the mini-filament, as we did with IRIS. However, a small number of profiles on the jet exhibit blueshifted Doppler velocities up to 6 km/s. On the other hand, we find redshifted values on the loop of up to 10 km/s.
Statistical analysis
In this section, we present correlations between Mg II k physical parameters, as well as between those and the Si IV derived from Paper I and refined as explained in paragraph 2.3.2. We used the RB asymmetry (introduced in 3.1.5), the peak separation, and the core depth. As mentioned in 2.3.1, we excluded Mg II k profiles with RL , as well as profiles with negative core depth (0.4%), resulting from fitting failures.
The peak separation (Pereira et al. [80]) is a measure of the difference between the peak wavelengths, defined as . The core depth (Schmit et al. [92]) describes the depth of the k3 intensity relative to the k2 peak intensities and : . A core depth equal to 1 means a profile with the deepest k3 absorption, while a core depth equal to 0 a profile that is not self-reversed. The above parameters were calculated for the loop and jet masks, as well as a background region for comparison, as defined in paragraph 3.1.5. We excluded the bright redshifted structure, as the Mg II k profiles there lack a clear central reversal, thus the calculation of the above parameters may be unreliable.
To quantify the relationships, we used the Spearman rank correlation coefficient, which measures the strength of monotonic associations between variables (Spearman [98]). This method is preferred over Pearson correlation in our analysis because the relationships between the parameters may be nonlinear or influenced by small sample sizes, and Spearman’s is more robust to these conditions.
Correlations of Mg II k parameters
Figure 8 presents the histograms of Mg II k parameters for each region: k3 intensity, peak separation, core depth and RB asymmetry. The kernel density estimate (KDE) (Rosenblatt [86]) is over-plotted, in order to give a smooth estimate of the underlying distribution. The bandwidth is determined by Scott’s rule7 (Scott [94]). Additionally, Figure 9 shows the scatter plots of the peak separation versus the core depth, RB asymmetry and k3 intensity. As previously mentioned, negative/positive RB asymmetry values correspond to blueshifts/redshifts.
Figure 8. Histograms with KDE overlay of Mg ii k parameters.
Figure 9. Scatter plots of Mg ii k parameters, with the Spearman correlation coefficients for all scatter points and each structure individually shown.
To assess differences in the distributions of the Mg II k parameters between regions, we applied the non-parametric Kruskal–Wallis H test (Kruskal & Wallis [52]), which evaluates whether the samples from the different regions are drawn from the same distribution without assuming that the data are normally distributed. Following, a Dunn’s post-hoc test (Dunn [30]) is applied for pairwise comparisons when significant (). The results are summarized in Table 1 and confirm statistically significant differences (denoted with in Table 1) across all measured parameters between the jet and the background, with the exception of the k3 intensity.
| Feature | BG-jet | BG-loop | jet-loop |
| Intensity | × | √ | √ |
| Peak Separation | √ | √ | √ |
| Core Depth | √ | × | √ |
| RB Asymmetry | √ | × | √ |
Table 1. Statistical comparison of spectral parameters.
For the region below the loop, the core depth and RB asymmetry show no significant differences from the background. However, this region exhibits higher k3 intensities and lower peak separations (see Figure 8). The jet is blueshifted, as indiNotes. Results are based on pairwise Dunn’s post-hoc tests. A checkmark (√) indicates a significant difference (), while a cross (×) indicates no significant difference. BG refers to the background.
cated by the RB asymmetry, with higher peak separations and core depths. We note that the mini-filament-like structure region, estimated roughly by its low intensity (green arrow in Figure 7) and the blueshifted stripe in Figure 6), is not statistically different from the background in any parameter, except for the intensity itself.
The peak separation is correlated with the core depth and the intensity for both structures (Figure 9). Correlations within individual structures are weaker, suggesting that the overall stronger correlations mainly reflect differences between the two structures. The correlations between the peak separation and RB asymmetry are almost zero for each structure separately. The slightly larger overall correlation highlights the distinction between the jet and the region below the loop and background, with the latter two being indistinguishable in this parameter. These relationships will be analyzed in detail in the Discussion (§ 4). We emphasize that these results should be interpreted with caution, as statistics do not necessarily imply physical relationships. However, they may reflect similar behavior at the atmospheric heights probed by these diagnostics.
Correlations between Mg ii k and Si iv 1393 Å parameters
Figure 10 shows the scatter plots between the Mg ii k parameters-peak separation, core depth, and RB asymmetry- and the Doppler velocities of the Si iv 1393 Å line for the core and tail components, which are derived from the bimodal Gaussian fitting performed to the Si iv profiles. The tail component is characterized as wider, with generally lower intensities and higher Doppler shifts, while the core component is narrower with higher intensities (see Paper I).
Figure 10. Scatter plots of the Si iv 1393 Å Doppler velocities of the core and tail components versus Mg ii k parameters. Different colors indicate the two distinct structures, as explained in the legend of panel A. The Spearman correlation coefficient for each panel is shown in the top-left corner for all points and each structure separately.
In all parameters, the jet generally presents higher correlations than the region below the loop. The correlations are higher when taking into account both structures together, an effect that mostly highlights the clear distinction between them in these parameters, rather than point-wise correlations. The jet is characterized by Mg ii k profiles with deep central reversals, large peak separations and blueshifted Si iv velocities, indicating the connection between the chromospheric and transition region signatures of the cool jet component. On the other hand, the loop exhibits redshifted Si iv velocities, associated with increased intensities and smaller peak separations of the Mg ii k profiles probing the region below it.
More specifically, the positive correlations between the core and tail Doppler velocities and the RB asymmetry in the case of the jet indicate similar plasma motions between the upper chromosphere and transition region. The zero correlation between the RB asymmetry and the Si iv 1393 Å tail component Doppler shifts within the loop region (panel f) suggests that the event causing the tail component profiles does not have an impact on the chromosphere, while the core component shows a stronger correlation. However this relationship may reflect a general behavior, since the loop is not distinguishable from the background in the RB asymmetry.
Generally, the three Mg ii k parameters were calculated by combining intensities originating from the middle chromosphere with the core intensity forming in the upper chromosphere. Because the middle chromosphere is only weakly coupled to the transition region, parameters that mix contributions from both layers may naturally exhibit weaker correlations with transition region diagnostics, typically below 50%. Moreover, the loop is not directly visible in upper chromospheric temperatures, as it is formed in a temperature range from the low transition region to the corona, in contrast to the jet, which includes a cool counterpart. Thus, the correlations for the loop are expected to be weaker or non-existent.
Discussion
In this section, we interpret our findings in the context of other studies in the literature and discuss their relevance to the formation and evolution of the observed jet.
Spectral signature
First of all, we observe a positive correlation between the peak separation and core depth (see paragraph 3.2.1), which was also found by Schmit et al. [92]. The enhancement of the peak separation can be explained by the effect of opacity broadening (Rathore & Carlsson [83]), where an increased opacity, corresponding to a higher column mass above the peaks formation height, results in a wider absorption profile, shifting the of the peaks to larger frequency offsets (Hansteen et al. [37]). The higher opacity may also raise to larger altitudes, where the source function is lower, so that the intensity is decreased, leading also to an increased core depth.
Especially on the jet spire, we find the largest peak separations and core depths, and the lowest intensities (Figure 8 and Figure 9). There, the upward plasma motions, measured in the RB asymmetry (Figure 7), may increase the column mass by transporting dense chromospheric plasma to larger heights, thereby enhancing the opacity. The variation in formation height is possibly implied by the observed redshifted velocities obtained by IRIS2 at the jet base for (Figure 6), as well as (not shown). Furthermore, the deeper central reversals, as well as the lower intensities on the jet spire (Figure 8), suggest an elevated height found in numerical simulations (see Leenaarts et al. [55]; Ondratschek et al. [71, 72] and Rathore et al. [84] for the C ii lines). While the relationships reported in the above studies are weak and show substantial scatter, they may still indicate an underlying trend.
Furthermore, the correlations of the Si iv velocities with both the RB asymmetry and the peak separation suggest that velocity gradients may provide an additional mechanism responsible for the enhanced peak separation. Simulations by Leenaarts et al. [55]; Ondratschek et al. [72] indicate that the peak separation is wider in the presence of larger velocity gradients, as the latter cause wavelength shifts in different heights. According to the same authors, an exception is the case of a wider absorption profile in the presence of a temperature maximum in the lower chromosphere. Another parameter that is believed to play a role in increasing the width of the absorption profile, thus enhancing the peak separation, is microturbulence (Carlsson et al. [17]; Hansteen et al. [37]). However, its exact nature and contribution to the formation of chromospheric spectral lines is still debated (Judge et al. [45]). Besides, in most cases models cannot fully reproduce the Mg ii h&k profiles and microturbulence is added as an ad hoc parameter (see Pereira et al. [80] and Bjørgen et al. (2018) for the Ca ii H&K lines).
Generally, the clear distinction between the region below the loop and the jet in all the above parameters demonstrates that the jet and loop leave distinct chromospheric and transition region signatures. The jet exhibits strong dynamical effects detectable in both Mg ii and Si iv, while the loop doesn’t influence the chromosphere in such extent, confirming that loops in general are coronal/transition region structures observed in higher temperatures.
Mini-filament-like structure evolution
Table 2 shows the timeline of events leading to the jet eruption. We observe a mini-filament-like structure (arrows A and B in Figure 3), which gets elongated and destabilized, possibly triggering the jet eruption. The mini-filament-like structure is initially observed very close to the jet base, similarly to other observations (Sterling et al. [99]).
Before the appearance of the mini-filament-like structure, we observe possible weak, partial flux cancellation at PIL1 (see Figure 4 and online HMI animation). Several recurrent increases in AIA 304 Å intensity are also observed above PIL1, presumably resulting from reconnection driven by cancellation. This phenomenon has been reported in similar settings early in the cancellation phase (Bura et al. [13]) or prior to the mini-filament formation (Panesar et al. [73]). Kontogiannis et al. [51] observe such increases in intensity before the mini-filament eruption, however without evidence of flux cancellation.
We also observe flux emergence and shear motions, which, along with flux cancellation, have been associated with the formation of magnetic arcades capable of supporting filament plasma, as they can increase the non-potentiality of the overlying magnetic configuration (van Ballegooijen & Martens [103]; Zhou [111]). Moreover, flux cancellation has also been associated with mini-filament eruptions, as it can facilitate the internal reconnection below it and lead to its destabilization (Panesar et al. [74]).
Upward motions and elongation of mini-filaments, similar to our mini-filament-like structure, have been reported by other studies, for example Panesar et al. [73] but in larger time-scales. The estimated lifetime of 5 to 8 minutes, given the uncertainty in the exact formation time, and the maximum length of 11 Mm, are both within the ranges reported for mini-filaments in previous studies (Sterling et al. [99]; Chen et al. [21]). Moreover, the mini-filament-like structure is partially sampled by the IRIS slit at its rising motion, seen as a faint structure with low intensities (see Figure 7) as well as blueshifted velocities of 4 km/s, low densities and temperatures derived from IRIS2 at (Figure 6).
The evolution of the mini-filament-like structure is connected with two distinct structures in different timestamps. First of all, the IRIS slit captures a bright redshifted structure, spatially coinciding with the mini-filament-like structure prior to its rising (Figure 3, Figure 6 and Figure 7). The high intensity seen in the Mg ii k spectra and 2796 Å SJI implies that it exhibits higher temperatures than its surroundings, in accordance with the higher temperature derived from IRIS2. The inversions also show increased electron densities with respect to the surrounding plasma, as well as redshifted velocities. Although the latter originate from a not well-constrained optical depth of , they are consistent with the redshift derived qualitatively from the RB asymmetry.
| Time (UT) | Event |
| 15:58 – 16:58 | Converging motions and decrease of magnetic flux at PIL1 |
| >17:17 | Appearance of the mini-filament-like structure |
| 17:20:53 – 17:23:31 | Downward flowing plasma detected from the southern end of the mini-filament-like structure |
| 17:23:43 | Upward motions and length increase of the mini-filament-like structure |
| 17:24:07 | Appearance of elongated bright structure in AIA 304 Å |
| 17:24:55 | Brightening in AIA 304 Å |
| 17:25:19 | Appearance of the cool jet counterpart in AIA 304 Å |
| 17:26:30 | Hot jet counterpart and loop first appearance in higher-temperature AIA channels |
Table 2. Characteristic events describing the mini-filament-like structure and jet evolution.
This structure may represent drainage of mini-filament material along magnetic field lines from its southern footpoint rooted in a region of negative magnetic polarity, associated with destabilization of the mini-filament. Kontogiannis et al. [51] report similar results, as downflows are detected at one footpoint of a mini-filament. Plasma drainage has also been observed and studied in large-scale filaments ([14] (Wu et al. 2025)), where plasma heating at the site of drainage is caused by transformation of kinetic to thermal energy ([32]).
Another possibility is that the bright redshifted structure is a result of magnetic reconnection between falling mini-filament plasma and pre-existing loops, assuming that the features marked by arrow D in Figure 3 correspond indeed to low-lying loops. In addition, the high temperature further supports the possibility of localized heating due to reconnection ([32]). Even if the mini-filament-like structure is not actually a mini-filament, the bright redshifted structure can still indicate plasma flowing downwards following the structure’s eruptive motion.
Nonetheless, all above scenarios are consistent with the fact that almost or fully single-peaked profiles dominate this region (Figure 2). Various studies show that single-peaked profiles can be produced in flares and plages due to increased densities and temperatures ([17, 48, 78]), as well as due to increased velocities at the line center ([87]). While the magnitude of these conditions differs from that in our observations, we do find increased temperatures and densities. Velocity gradients can also alter the shape of the profiles to such extent that the red peak is almost extinct, leading to an absent central reversal. To our knowledge, our work is one of the few studies observing such profiles associated with mini-filaments and jets.
The second structure associated with the evolution of the mini-filament-like structure is an elongated feature, visible in AIA 304 Å. It forms parallel to the mini-filament, develops concurrently with its rise (Figure 3, arrow E1) and probably consists of multiple smaller-scale structures (Figure 3, arrow E2). A similar elongated bright structure was reported by Madjarska et al. [65], who interpreted it as newly reconnected loops developing above the PIL and associated with quasi-separatrix layer (QSL) reconnection during the evolution of a micro-flare. In our case, it may represent loops associated with the destabilization and eruption of the mini-filament, forming after internal reconnection below it (Li et al. 2023), in the case that a mini-filament is indeed present. However, we emphasize that this interpretation remains speculative.
Magnetic topology and reconnection
Generally, the potential field extrapolation provides evidence that magnetic reconnection may play a role in the mini-filament-like structure evolution and jet triggering (Figure 5), as we find a null point at a height of 2 Mm in close proximity to the mini-filament’s northern edge and a QSL close to the base of the jet between opposite sign polarities, serving as a favorable location for reconnection ([36] (Li et al. 2025)). Null point reconnection has been associated with the eruption of mini-filaments and jets ([53] (Cheng et al. 2023)), as well as large-scale filaments ([29] (Li et al. 2017)). Additional indications of reconnection at the jet base include redshifted velocities derived from the IRIS2 inversion (Figure 6-arrow A), as well as the appearance of a brightening in AIA 304 Å at the same location before the jet eruption (see online animation).
The jet consists of two components, with a cooler counterpart preceding the hotter one by approximately 1 to 2 minutes (arrows F1 and F2 in Figure 3, respectively). Hot jet counterparts are often observed before the cooler ones ([43]), with the latter associated with mini-filament eruptions ([69]). However, Shen et al. [96] report the opposite sequence, in which the cooler component appeared before the hotter counterpart. They suggest either that reconnection occurred at lower chromospheric heights or that the hot plasma resulted from heating of the cooler material. Both interpretations are broadly consistent with our observations, since we detect signatures of the jet in chromospheric diagnostics and a magnetic null point at the upper chromosphere.
A possible mechanism for the formation of the jet and the associated loop is interchange reconnection between the rising mini-filament-like structure and the overlying open magnetic field lines ([99, 75]). The potential field extrapolation topology reveals closed magnetic field lines spatially coincident with the loop and open magnetic field lines above the rising mini-filament-like structure, thereby providing a configuration favorable for this scenario.
Conclusion
In this work, we presented a comprehensive analysis of a solar jet, within a low-latitude coronal hole, combining diagnostics from the Mg II k and Si IV spectral lines. By merging IRIS and SDO observations, we achieved a multi-layer study of the jet. The spectral analysis of Mg II k and Si IV spectra diagnostics revealed a coupling between the chromospheric and transition region dynamics and plasma motions. Although the jet is not directly visible in the Mg ii k intensity and the SJ 2796 Å images, there is a signature of the jet in the shape and behavior of the Mg ii k profiles. The increased opacity caused by the jet plasma results in increased peak separations and core depths of the Mg ii k profiles.
The jet eruption is also associated with a small-scale absorption structure, most likely a mini-filament, whose formation is preceded by magnetic flux emergence, shear motions, and weak flux cancellation in the photosphere. The mini-filament-like structure follows upward motions, accompanied by downward flows at its footpoint, possibly due to drainage. There, we find almost or fully single-peaked Mg ii k profiles as a result of enhanced thermodynamic conditions or plasma motions.
Although the exact mechanism responsible for the jet eruption cannot be fully elucidated, reconnection likely plays a role in triggering the evolution of the system, evident by the presence of a null point roughly at the upper chromosphere and a QSL below the jet base. The appearance of the cooler jet component earlier than the hotter one, along with the location of the null point and the chromospheric response of all structures associated with the jet eruption, leads us to the conclusion that the origin of the jet can be traced down to the chromosphere.
Overall, this study provides insight into the genesis of the jet and the chromosphere – transition region connection. A novelty of this work is the use of GA and RL for the fitting of the Mg ii and Si iv profiles, making the fits more robust against noise and spectral profile variability. The expansion of this work to more observations is key in determining the generality of the above conclusions. Unfortunately, the size of the structures under study is close to the resolution limit of the related instruments, making the analysis of this kind challenging. Nevertheless, such detailed multiwavelength observations of this kind can pose a significant challenge for MHD models of jets.
Acknowledgements. We thank the anonymous referee for insightful comments and suggestions that improved this work. We would also like to thank J.-C. Vial for useful discussions. C.G. acknowledges support by the Sectoral Development Program (ΟΠΣ 5223471) of the Ministry of Education, Religious Affairs and Sports, through the National Development Program (NDP) 2021-25. S.P. acknowledges support by the ERC Synergy Grant ‘Whole Sun’ (GAN: 810218).
References
- [1]Adams, M., Sterling, A. C., Moore, R. L., & Gary, G. A. 2014, ApJ, 783, 11
- [2]Akaike, H. 1974, IEEE Transactions on Automatic Control, 19, 716
- [3]Alexander, D. & Fletcher, L. 1999, Sol. Phys., 190, 167
- [4]Alissandrakis, C. E. 1981, A&A, 100, 197
- [5]Antolin, P., Auchère, F., Winch, E., Soubrié, E., & Oliver, R. 2024, Sol. Phys., 299, 94
- [6]Anzer, U. & Heinzel, P. 2005, ApJ, 622, 714
- [7]Archontis, V. & Syntelis, P. 2019, Philosophical Transactions of the Royal Society of London Series A, 377, 20180387
- [9]Athay, R. G. & Skumanich, A. 1968, Sol. Phys., 3, 181
- [10]Barczynski, K., Schmieder, B., Gelly, B., Peat, A. W., & Labrosse, N. 2023, A&A, 680, A63
- [11]Böse, S., Henriques, V. M. J., Rouppe van der Voort, L., & Pereira, T. M. D. 2019, A&A, 627, A46
- [12]Brueckner, G. E. & Bartoe, J.-D. F. 1983, ApJ, 272, 329
- [13]Bura, A., Samanta, T., Prasad, A., et al. 2025, ApJ, 985, L47
- [14]Cabezas, D. P., Ichimoto, K., Asai, A., et al. 2024, A&A, 690, A172
- [16]Canfield, R. C., Reardon, K. P., Leka, K. D., et al. 1996, ApJ, 464, 1016
- [17]Carlsson, M., Leenaarts, J., & De Pontieu, B. 2015, ApJ, 809, L30
- [18]Carlsson, M. & Stein, R. F. 1997, ApJ, 481, 500
- [19]Cartwright, K. 2017, Journal of Mathematical Science and Mathematics Education, 12
- [20]Chae, J., Qiu, J., Wang, H., & Goode, P. R. 1999, ApJ, 513, L75
- [21]Chen, H., Tian, H., Zhang, Q., et al. 2025, ApJ, 983, 143
- [22]Chen, H., Yang, J., Yang, B., Ji, K., & Bi, Y. 2018, Sol. Phys., 293, 93
- [24]Chitta, L. P., Huang, Z., D’Amicis, R., et al. 2025, A&A, 694, A71
- [25]da Silva Santos, J. M., White, S. M., Reardon, K., et al. 2022, Frontiers in Astronomy and Space Sciences, 9, 898115
- [26]de la Cruz Rodríguez, J., Leenaarts, J., Danilovic, S., & Uitenbroek, H. 2019, A&A, 623, A74
- [27]De Pontieu, B., Title, A. M., Lemen, J. R., et al. 2014, Sol. Phys., 289, 2733
- [28]Del Zanna, G., Dere, K. P., Young, P. R., & Landi, E. 2021, The Astrophysical Journal, 909, 38
- [29]Devi, P., Mandrini, C. H., Chandra, R., et al. 2025, Sol. Phys., 300, 163
- [30]Dunn, O. J. 1964, Technometrics, 6, 241
- [31]Fortin, F.-A., De Rainville, F.-M., Gardner, M.-A., Parizeau, M., & Gagné, C. 2012, Journal of Machine Learning Research, 13, 2171
- [32]Gilbert, H. R., Inglis, A. R., Mays, M. L., et al. 2013, ApJ, 776, L12
- [33]Gontikakis, C. & Vial, J.-C. 2018, A&A, 619, A64
- [34]Gorman, J., Chitta, L. P., & Peter, H. 2022, A&A, 660, A116
- [35]Gouttebroze, P. 1989, ApJ, 337, 536
- [36]Guo, Y., Démoulin, P., Schmieder, B., et al. 2013, A&A, 555, A19
- [37]Hansteen, V. H., Martinez-Sykora, J., Carlsson, M., et al. 2023, ApJ, 944, 131
- [38]Haynes, A. L. & Parnell, C. E. 2007, Physics of Plasmas, 14, 082107
- [39]Heinzel, P., Vial, J.-C., & Anzer, U. 2014, A&A, 564, A132
- [40]Holland, J. H. 1992, Scientific American, 267, 66
- [41]Hong, J., Jiang, Y., Zheng, R., et al. 2011, ApJ, 738, L20
- [42]Innes, D. E., Inhester, B., Axford, W. I., & Wilhelm, K. 1997, Nature, 386, 811
- [43]Jiang, Y. C., Chen, H. D., Li, K. J., Shen, Y. D., & Yang, L. H. 2007, A&A, 469, 331
- [45]Judge, P. G., Kleint, L., Leenaarts, J., Sukhorukov, A. V., & Vial, J.-C. 2020, ApJ, 901, 32
- [46]Karki, G., Guo, J., Schmieder, B., et al. 2026, ApJ, 999, 148
- [47]Kass, R. E. & Raftery, A. E. 1995, Journal of the American Statistical Association, 90, 773
- [48]Kerr, G. S., Simões, P. J. A., Qiu, J., & Fletcher, L. 2015, A&A, 582, A50
- [49]Kneer, F., Scharmer, G., Mattig, W., et al. 1981, Sol. Phys., 69, 289
- [50]Koletti, M., Gontikakis, C., Patsourakos, S., & Tsinganos, K. 2024, A&A, 690, A11
- [51]Kontogiannis, I., Dineva, E., Diercke, A., et al. 2020, ApJ, 898, 144
- [52]Kruskal, W. H. & Wallis, W. A. 1952, Journal of the American Statistical Association, 47, 583
- [53]Kumar, P., Karpen, J. T., Antiochos, S. K., et al. 2019, ApJ, 873, 93
- [54]Leenaarts, J., Pereira, T. M. D., Carlsson, M., Uitenbroek, H., & De Pontieu, B. 2013a, ApJ, 772, 89
- [55]Leenaarts, J., Pereira, T. M. D., Carlsson, M., Uitenbroek, H., & De Pontieu, B. 2013b, ApJ, 772, 90
- [56]Lemaire, P., Choucq-Bruston, M., & Vial, J.-C. 1984, Sol. Phys., 90, 63
- [57]Lemaire, P., Gouttebroze, P., Vial, J. C., & Artzner, G. E. 1981, A&A, 103, 160
- [58]Lemen, J. R., Title, A. M., Akin, D. J., et al. 2012, Sol. Phys., 275, 17
- [62]Liu, Y., Hoeksema, J. T., Scherrer, P. H., et al. 2012, Sol. Phys., 279, 295
- [63]Livi, S. H. B., Wang, J., & Martin, S. F. 1985, Australian Journal of Physics, 38, 855
- [65]Madjarska, M. S., Mackay, D. H., Galsgaard, K., Wiegelmann, T., & Xie, H. 2022, A&A, 660, A45
- [66]Martin, S. F. 1988, Sol. Phys., 117, 243
- [67]Milkey, R. W. & Mihalas, D. 1974, ApJ, 192, 769
- [69]Moore, R. L., Cirtain, J. W., Sterling, A. C., & Falconer, D. A. 2010, ApJ, 720, 757
- [71]Ondratschek, P., Przybylski, D., Smitha, H. N., et al. 2024, A&A, 692, A6
- [72]Ondratschek, P., Przybylski, D., Smitha, H. N., et al. 2026, A&A, 711, A216
- [73]Panesar, N. K., Sterling, A. C., & Moore, R. L. 2017, ApJ, 844, 131
- [74]Panesar, N. K., Sterling, A. C., & Moore, R. L. 2018, ApJ, 853, 189
- [75]Panesar, N. K., Sterling, A. C., Moore, R. L., & Chakrapani, P. 2016, ApJ, 832, L7
- [77]Panesar, N. K., Tiwari, S. K., Moore, R. L., Sterling, A. C., & De Pontieu, B. 2022, ApJ, 939, 25
- [78]Panos, B., Kleint, L., Huwyler, C., et al. 2018, ApJ, 861, 62
- [79]Patsourakos, S. & Archontis, V. 2025, A&A, 699, A87
- [80]Pereira, T. M. D., Leenaarts, J., De Pontieu, B., Carlsson, M., & Uitenbroek, H. 2013, ApJ, 778, 143
- [82]Priest, E. R. & Démoulin, P. 1995, J. Geophys. Res., 100, 23443
- [83]Rathore, B. & Carlsson, M. 2015, The Astrophysical Journal, 811, 80
- [84]Rathore, B., Carlsson, M., Leenaarts, J., & Pontieu, B. D. 2015, The Astrophysical Journal, 811, 81
- [85]Régnier, S. 2012, Sol. Phys., 277, 131
- [86]Rosenblatt, M. 1956, The Annals of Mathematical Statistics, 27, 832
- [87]Rubio da Costa, F. & Kleint, L. 2017, ApJ, 842, 82
- [88]Sainz Dalda, A., de la Cruz Rodríguez, J., De Pontieu, B., & Gošić, M. 2019, ApJ, 875, L18
- [89]Sainz Dalda, A. & De Pontieu, B. 2023, Frontiers in Astronomy and Space Sciences, 10, 1133429
- [90]Savitzky, A. & Golay, M. J. E. 1964, Analytical Chemistry, 36, 1627
- [91]Scherrer, P. H., Schou, J., Bush, R. I., et al. 2012, Sol. Phys., 275, 207
- [92]Schmit, D., Bryans, P., De Pontieu, B., et al. 2015, ApJ, 811, 127
- [93]Schwarz, G. 1978, Annals of Statistics, 6, 461
- [94]Scott, D. W. 1992, in Wiley Series in Probability and Statistics
- [95]Shen, Y. 2021, Proceedings of the Royal Society of London Series A, 477, 217
- [96]Shen, Y., Liu, Y. D., Su, J. Q., Zu, X., & Tian, Z. 2017, The Astrophysical Journal, 851, 67
- [98]Spearman, C. 1904, The American Journal of Psychology, 15, 72
- [99]Sterling, A. C., Moore, R. L., Falconer, D. A., & Adams, M. 2015, Nature, 523, 437
- [100]Titov, V. S., Hornig, G., & Démoulin, P. 2002, Journal of Geophysical Research (Space Physics), 107, 1164
- [103]van Ballegooijen, A. A. & Martens, P. C. H. 1989, ApJ, 343, 971
- [104]Wang, J., Li, W., Denker, C., et al. 2000, The Astrophysical Journal, 530, 1071
- [105]Wilks, S. S. 1938, Annals of Mathematical Statistics, 9, 60
- [106]Williams, D. R., Baker, D., & van Driel-Gesztelyi, L. 2013, ApJ, 764, 165
- [107]Wright, S. 1921, Journal of agricultural research, 20, 557
- [110]Zhang, P., Chen, J., Liu, R., & Wang, C. 2022, ApJ, 937, 26
- [111]Zhou, Y. 2025, Reviews of Modern Plasma Physics, 9, 32
- [112]Zhu, Y., Kowalski, A. F., Tian, H., et al. 2019, ApJ, 879, 19
- [113]Zhuleku, J., Archontis, V., & Moraitis, K. 2025, ApJ, 986, 47







