The Astrophysical Journal, 930:129 (11pp), 2022 May 10 https://doi.org/10.

3847/1538-4357/ac5c53
© 2022. The Author(s). Published by the American Astronomical Society.

High-frequency Waves in Chromospheric Spicules

W. Bate1 , D. B. Jess1,2 , V. M. Nakariakov3,4 , S. D. T. Grant1 , S. Jafarzadeh5,6 , M. Stangalini7 , P. H. Keys1 ,
D. J. Christian2 , and F. P. Keenan1
1
Astrophysics Research Centre, School of Mathematics and Physics, Queen’s University Belfast, Belfast, BT7 1NN, UK; wbate02@qub.ac.uk
2
Department of Physics and Astronomy, California State University Northridge, 18111 Nordhoff Street, Northridge, CA 91330, USA
3
Centre for Fusion, Space and Astrophysics, Physics Department, University of Warwick, Coventry, CV4 7AL, UK
4
St. Petersburg Branch, Special Astrophysical Observatory, Russian Academy of Sciences, 196140, St. Petersburg, Russia
5
Leibniz Institute for Solar Physics (KIS), Schöneckstr. 6, D-79104 Freiburg, Germany
6
Rosseland Centre for Solar Physics, University of Oslo, P. O. Box 1029 Blindern, NO-0315 Oslo, Norway
7
ASI, Italian Space Agency, Via del Politecnico snc, I-00133, Rome, Italy
Received 2022 February 4; revised 2022 February 24; accepted 2022 March 8; published 2022 May 11

Abstract
Using high-cadence observations from the Hydrogen-alpha Rapid Dynamics camera imaging system on the Dunn
Solar Telescope, we present an investigation of the statistical properties of transverse oscillations in spicules
captured above the solar limb. At five equally separated atmospheric heights, spanning approximately 4900–7500
km, we have detected a total of 15,959 individual wave events, with a mean displacement amplitude of
151 ± 124 km, a mean period of 54 ± 45 s, and a mean projected velocity amplitude of 21 ± 13 km s−1. We find
that both the displacement and velocity amplitudes increase with height above the solar limb, ranging from
132 ± 111 km and 17.7 ± 10.6 km s−1 at ≈4900 km, and 168 ± 125 km and 26.3 ± 14.1 km s−1 at ≈7500 km,
respectively. Following the examination of neighboring oscillations in time and space, we find 45% of the waves to
be upwardly propagating, 49% to be downwardly propagating, and 6% to be standing, with mean absolute phase
velocities for the propagating waves on the order of 75–150 km s−1. While the energy flux of the waves
propagating downwards does not appear to depend on height, we find the energy flux of the upwardly propagating
waves decreases with atmospheric height at a rate of −13,200 ± 6500 W m−2/Mm. As a result, this decrease in
energy flux as the waves propagate upwards may provide significant thermal input into the local plasma.
Unified Astronomy Thesaurus concepts: Solar spicules (1525); Solar oscillations (1515); Solar chromosphere
(1479); Solar atmosphere (1477)

1. Introduction One of the major focuses of current solar physics research is

the so-called “coronal heating paradox.” One of the proposed
Spicules are dynamic plasma jets that are prevalent within the
theoretical mechanisms to explain the source of this heating is
solar chromosphere and that generally have diameters on the
linked to the propagation and dissipation of wave phenomena,
order of hundreds of kilometers. They are relatively short-lived
commonly referred to as the “AC” heating mechanism (Alfvén
features, typically having a lifetime of less than 10 minutes
& Lindblad 1947). Spicules are of particular interest when
(Pereira et al. 2012). When viewed in the visible and UV bands
attempting to explain the heating of the solar atmosphere due to
at the solar limb, spicules appear ubiquitously as a dense forest
their potential to facilitate the transfer of mass and energy
of narrow, straw-like features (Sterling 1998, 2000).
Secchi (1875) was the first to observe solar spicules, and between the photosphere and corona. They are often categor-
they have remained a focal point of solar physics research over ized by their properties into two types, type I and type II (De
the last 140 yr. Transverse oscillations in spicules were first Pontieu et al. 2007a; Pereira et al. 2012), although such distinct
identified in the 1960s (Pasachoff et al. 1968), utilizing ground- classifications are still under debate (e.g., Zhang et al. 2012;
based observations obtained at the Sacramento Peak Observa- Pereira et al. 2013).
tory (for a comprehensive review, see Zaqarashvili & Observations of transverse oscillations of spicules, fibrils,
Erdélyi 2009). The magnetic cylinder model is generally and mottles in the upper chromosphere typically find mean
accepted as being the most applicable to spicules, allowing periods on the order of 80–300 s, often with ∼50–1000
their oscillatory behavior to be interpreted and modeled as examples found during the course of an individual data
magnetohydrodynamic (MHD; Alfvén 1942) modes (Edwin & sequence (Nikol’Skii & Sazanov 1967; Nikolsky & Platova
Roberts 1983). Sterling (2000) highlighted that high-resolution 1971; De Pontieu et al. 2007b; Kuridze et al. 2012; Morton
observations, due to the small width of the structures, are vital et al. 2012, 2013, 2014). A major exception are the short-period
for a complete description of the spicule wave phenomena. (45 ± 30 s) transverse oscillations found in 89 type II spicules
Wedemeyer-Böhm et al. (2007) also note that the ability to within a coronal hole identified by Okamoto & De Pontieu
detect oscillatory power at higher frequencies is influenced by (2011). However, Okamoto & De Pontieu (2011) suggest that
the spatial resolution of the observations (see also the this short average period is likely due to them utilizing a
discussions provided by Jess et al. 2020, 2021). methodology that did not allow for the measurement of the
properties of the longer period waves (>100 s). Another
Original content from this work may be used under the terms
exception is the more recent work of Shetye et al. (2021),
of the Creative Commons Attribution 4.0 licence. Any further which found 30 examples of transverse spicule oscillations
distribution of this work must maintain attribution to the author(s) and the title with periods ranging from 16 to 100 s. These authors also note
of the work, journal citation and DOI. a selection effect due to only choosing an event if its oscillation

1
The Astrophysical Journal, 930:129 (11pp), 2022 May 10 Bate et al.

period is less than its lifetime, resulting in longer period waves significant amount of wave energy may be unaccounted for
not being considered. (Verth & Jess 2016). Another aspect contributing to the
Finding the energy flux of spicule oscillations is an underestimation of the total energy flux may be the presence of
important step in investigating their contribution to the heating kink motions along the observer’s line of sight, which will not
of the chromosphere and corona. It is estimated that an energy manifest as visible transverse oscillations. Examples of this
flux of 103–104 W m−2 is required to heat the chromosphere. have been documented by Sharma et al. (2018) and Shetye
The energy required to heat the corona is around an order of et al. (2021), who measured helical motions of spicules through
magnitude less than that required to heat the chromosphere Doppler measurements (see also the modeling work by
(Withbroe & Noyes 1977). This suggests that accounting for Zaqarashvili & Skhirtladze 2008). If these line-of-sight motions
chromospheric heating is a challenge of equal or greater are not taken into account when calculating the energy flux, it
magnitude than for coronal heating when investigating solar may result in an underestimation of the true value.
atmospheric heating mechanisms. The aim of the current study is to identify the properties of
Energy flux estimations are based on the interpretation of spicule oscillations across a statistically significant sample that
these transverse oscillations as MHD wave modes. De Pontieu is extracted from different chromospheric heights. With
et al. (2007b) interpreted the transverse oscillations of spicules oscillation characteristics measured across a range of atmo-
as bulk Alfvén waves and assumed a filling factor of unity. spheric layers, we calculate the energy flux carried by these
Using this interpretation, an energy flux estimate of 4000– waves as a function of geometric height. To achieve this
7000 W m−2 was suggested. However, this bulk Alfvén objective, we utilize ground-based instrumentation with high
interpretation has attracted criticism, with Erdélyi & Fedun spatial and temporal resolutions, providing unprecedented data
(2007) and Van Doorsselaere et al. (2008) pointing out that products that are ideally suited for this study.
Alfvén waves do not result in the bulk transverse motions
observed, and instead proposing that these transverse oscilla- 2. Observations
tions are best interpreted as kink modes. The filling factor, a
measure of what fraction of the total volume is occupied by Our analysis employs data collected on 2015 July 27 from
oscillating spicules, is another extremely important considera- 13:52–15:29 UT using the Dunn Solar Telescope (DST;
tion for energy flux calculations (Van Doorsselaere et al. 2014). Dunn 1969) at the National Solar Observatory in New Mexico,
An equivalent interpretation, assuming that spicules have an USA. The Rapid Oscillations in the Solar Atmosphere (ROSA;
Jess et al. 2010a) and Hydrogen-alpha Rapid Dynamics camera
approximately constant width across varying heights, would be
(HARDcam; Jess et al. 2012a) imaging systems were used to
the ratio of the area of the solar surface covered by spicules to
observe a large sunspot, which was part of NOAA AR12391,
the total solar surface area.
close to the solar limb at N07.8E73.6 in the conventional
Makita (2003) found a spicule filling factor of 5% at a height
heliographic coordinate system. Seeing conditions remained
of 4000 km using Ca II H and K line observations taken during
excellent throughout the first hour of the observing period,
a solar eclipse. This suggests that a filling factor of 0.05 is more
gradually worsening toward the latter stages of the observing
appropriate than unity. Using the revised interpretations of the window.
most realistic MHD mode and associated filling factor, Van HARDcam observations employed a narrow (0.25 Å FWHM)
Doorsselaere et al. (2014) found that the energy flux estimates bandpass filter centered on the Hα line core (6562.8 Å), while
by De Pontieu et al. (2007b) were reduced from 4000– the ROSA camera system observed the same region through
7000 W m−2 to 200–700 W m−2, a difference exceeding one G-band (10 Å FWHM centered at 4305 Å) and broadband
order of magnitude. Furthermore, Morton et al. (2012) used 4170 Å continuum filters. The HARDcam data have a pixel
high-resolution Hα observations taken by the Dunn Solar size of 0 092 (66.5 km), providing a 180″ × 180″ field of view,
Telescope to find a similar upper limit for the filling factor while the ROSA system was slightly undersampled (0 180 per
(4%–5%) for open chromospheric structures that connected to pixel) to provide an identical field-of-view size to that of the
higher layers of the solar atmosphere. By interpreting the HARDcam observations. To correct for wave-front deformations
transverse oscillations of fibrils as kink modes, the authors in real time, higher order adaptive optics (AO) were used during
estimated the energy flux as 170 ± 110 W m−2, similar to that the observations (Rimmele 2004; Rimmele & Marino 2011).
derived by Van Doorsselaere et al. (2014). However, in Original data from both ROSA and HARDcam were taken at a
addition to the 4%–5% filling factors commonly used in frame rate of 30.3 s−1, with the images synchronized by way of a
modern literature, lower estimates have also been put forward, master trigger with microsecond precision (Jess et al. 2010a).
with Beckers (1972) suggesting a filling factor of 0.6%. As a The resulting HARDcam Hα images were then improved using
result, it is generally believed that the spicule filling factor speckle reconstruction algorithms (Wöger et al. 2008), utilizing a
spans an approximate order of magnitude (ranging between 30 → 1 restoration, resulting in a final reconstructed cadence of
≈0.5% and 5%), with differing values being applicable 0.990 s.
depending on factors such as the atmospheric height sampled ROSA continuum observations were coaligned using cross-
and the degree of solar activity (i.e., it is not a quantity that can correlation techniques (see, e.g., Jess et al. 2010b) with
be applied universally across all observations). The influence of contemporaneous continuum images from the Helioseismic and
the chosen filling factor on energy flux calculations is discussed Magnetic Imager (HMI; Schou et al. 2012) on board the Solar
further in Section 3. Dynamics Observatory (SDO; Pesnell et al. 2012), providing
An important caveat when interpreting these energy flux subarcsecond pointing accuracy for the field of view covered
estimates is that they are only based on resolved transverse by the DST. Following this, the HARDcam field was aligned
oscillations. Waves with amplitudes too small to be spatially with the master ROSA images using sequences of targets
resolved or periods too short to be temporally resolved are not acquired during the calibration procedures at the DST, resulting
included in these estimations, leaving the possibility that a in Hα observations that have precise pointing metadata that is

2
The Astrophysical Journal, 930:129 (11pp), 2022 May 10 Bate et al.

Figure 1. Contextual SDO/HMI continuum (left), ROSA G-band (middle), and Hα line core (right) images acquired at 13:59:09 UT. The area imaged by ROSA and
HARDcam is marked by the red square in the full disk SDO/HMI continuum image. Numerous spicules are clearly visible above the solar limb as narrow, straw-like
structures in the corresponding Hα image.

consistent with modern space-based observatories. Contextual production of gamma-transformed images with a γ value of 3.2
images from SDO/HMI, ROSA, and HARDcam, following the (Poynton 2003).
processing steps outlined above, are shown in Figure 1. Five slits were placed at equally spaced, constant radial
heights above the limb, spanning approximately 4900 km to
7500 km in steps of ≈650 km. These slits were curved in
3. Analysis and Discussion nature in order to maintain a constant radial height above the
limb, and the highest and lowest slits are shown by the white
During the course of the observations, the DST’s AO system
lines in Figure 2. When taking this approach, it is important to
was locked onto the high-contrast sunspot structure that was
note that superposition along the line of sight of spicules
very close to the limb. As a result, limb spicules close to the
central portion of the field of view were accurately corrected anchored behind the limb, in front of the limb, and on the limb
from atmospheric seeing effects by the AO. Hence, the current is unavoidable. As a result of the slit heights being based on a
HARDcam Hα data set offers an unprecedented opportunity to geometric distance above the limb, this will result in the
examine limb spicules at extremely high time cadence (0.990 s) foreground/background spicules being sampled further along
and spatial resolution (133 km two-pixel resolution). their lengths than those precisely located on the limb. We have
A subfield, spanning approximately 70 Mm along the central carefully selected the minimum and maximum heights of the
portion of the field of view, where the AO corrections were slits to be in the range of 4890–7500 km (see Figure 2), which
operating optimally, was isolated for further study. As the DST is toward the upper end of the “dense forest” of spicules, hence
was tracking the sunspot contained within the field of view, minimizing the degree of feature superposition. Due to the
over time the pixel coordinates corresponding to the limb (minimized) spicule superposition affecting each of the slits in
position change as a result of the sunspot rotating onto the disk. a similar way, and considering the large numbers of spicules
The image sequence was hence stabilized with respect to the observed at each height, comparisons between wave properties
limb, which was achieved by first choosing a reference frame taken with different slits will still be valid. However, it is still
toward the beginning of the data set. Next, a small area of the important to consider this effect when examining wave
limb image with high contrast was selected, with subsequent properties taken from a single slit in isolation, since the chosen
images compared and shifted using two-dimensional cross- slit height will be a minimum value of the distance sampled
correlation techniques. Pixel shift values that produced the along the spicule due to these geometric considerations. Time–
highest cross-correlation coefficients were selected and applied distance diagrams were then produced from each of the slits,
to each image in the time series iteratively. The resulting with an example shown in Figure 3.
shifted images lead to the limb remaining stationary at the same The Automatic Northumbria University Wave Tracking
pixel location throughout the data set, providing a robust (Auto-NUWT; Morton et al. 2013; Weberg et al. 2018) code
baseline from which to examine spicule oscillations above the was utilized in order to identify the location of the spicules as a
fixed limb. function of time, track their transverse motion, and extract the
Multiscale Gaussian Normalization (MGN; Morgan & properties of their oscillations. Features are identified by fitting
Druckmüller 2014) was applied to each image in the data set a sum of Gaussian curves to each time slice in the time–
in order to more easily identify each spicule and its associated distance diagrams, enabling the determination of subpixel
motion. It must be noted that MGN does not preserve values for the location of the center of the feature. The
photometric accuracy. However, this is not an issue when transverse oscillatory behavior of these features is probed
mapping the transverse oscillations of features since we are not through the application of Fourier analysis to the position of the
concerned with comparisons of relative intensities. For the center of the feature as a function of time.
application of MGN, we employed the convolution of At each of the five heights considered, over 3000 spicule
HARDcam images with Gaussian kernels with one-sigma features are detected in the time–distance diagrams. Employing
widths of w = 1.25, 2.5, 5, 10, 20, 40 pixels, followed by the Fourier analysis, the properties of the waves present in the

3
The Astrophysical Journal, 930:129 (11pp), 2022 May 10 Bate et al.

Figure 2. An Hα core subfield (67 × 16 Mm2) image acquired using

HARDcam at 14:49:45 UT. Numerous spicules are clearly visible above the
solar limb as narrow, straw-like structures. The two most extreme slits used to
take the time–distance diagrams are shown by the white lines, at heights of
4890 and 7500 km. The axes are shown using different scales to aid with visual
clarity.

Figure 4. Histograms of the wave properties identified at a height of 6850 km

above the solar limb. The upper, middle, and lower panels display information
related to the displacement amplitudes, periods, and velocity amplitudes,
respectively. Measurements of the corresponding averages and deviations are
displayed on the right of each histogram. MAD denotes the median absolute
deviation.

Table 1
Mean wave Properties and their Standard Deviations at Each Sampled Height

Number of Displacement Velocity

Height Waves Amplitude Period Amplitude
(km) (km) (s) (km s−1)
4890 4880 132.4±111.2 55.1±45.0 17.7±10.6
5550 4920 139.6±118.8 57.0±48.1 18.3±10.7
6200 5022 152.4±128.5 55.8±48.1 20.6±11.6
6850 5209 163.7±128.6 53.4±45.4 23.4±13.1
7500 5298 168.1±125.4 48.9±39.9 26.3±14.1

distributions for these properties are consistent with those

found in previous studies (e.g., De Pontieu et al. 2007b;
Okamoto & De Pontieu 2011; Pereira et al. 2012).
Figure 3. A time–distance diagram captured using a curved slit at a height of Average wave properties for each height are presented in
6850 km above the solar limb. Each bright streak is a feature passing through Table 1. Importantly, the averages of the displacement and
the slit, with the clear oscillatory features representative of transverse motions velocity amplitudes appear to be consistent with those found in
displaying a range of amplitudes and periods. previous studies of transverse waves in spicules. However, the
average period of the waves in the current study are shifted to
lower values than those found previously (see the summary
transverse motions of these features were determined. As a provided by Jess et al. 2015). Specifically, the majority of
representative example, the averages and deviations of wave earlier studies found average periods on the order of 80–300 s
properties found at a height of 6850 km above the limb are (e.g., Nikolsky & Platova 1971; De Pontieu et al. 2007b;
displayed in Figure 4, where the distributions of the Kuridze et al. 2012; Morton et al. 2012, 2013, 2014), while we
displacement amplitudes, periods, and calculated velocity find the average period to be 53 ± 45 s (middle panel of
amplitudes of these waves are plotted as histograms. These Figure 4). It should be noted that the mean has been chosen for
properties follow approximate log-normal distributions, which comparison with previous studies here. However, as the wave
are shown by the solid green lines in Figure 4. Log-normal properties approximately follow log-normal distributions, the

4
The Astrophysical Journal, 930:129 (11pp), 2022 May 10 Bate et al.

Figure 6. Mean values of spicule displacement amplitudes (top panel) and

Figure 5. The displacement curves, corresponding to two of the waves velocity amplitudes (lower panel) plotted against height above the solar limb.
identified in our data set, are shown (in their raw form) using the solid black Linear lines of best fit are shown in both panels using a dashed red line. In the
lines. The dashed red lines highlight the fitted waves with properties derived bottom panel, a quadratic fit is shown using a dashed blue line. Errors
using Fourier analysis. The top panel shows a wave with a period of ≈138 s associated with each data point represent the standard error on the mean.
and a displacement amplitude of ≈358 km, while the bottom panel shows a
wave with a period of ≈20 s and a displacement amplitude of ≈79 km.
of Figure 5 has properties consistent with those found in
modal value represents a more useful statistic in understanding previous studies of transverse spicule oscillations (see the
the peak of this distribution. As a point of comparison, review by Jess et al. 2015), highlighting that these longer
Nikolsky & Platova (1971) found modal and mean periods of period (>50 s) waves are also present within our data and are
60 s and 85 s, respectively, for their observed spicule oscilla- fitted well using our techniques. However, due to the high
tions, whereas example modal and mean periods found at a spatial and temporal resolutions provided by HARDcam, much
height of 6850 km above the solar limb in this study are ≈24 s shorter period waves are able to be identified, including the
and ≈54 s, respectively, as shown in Figure 4. We consider the example shown in the lower panel of Figure 5.
detection of these shorter period waves likely due, at least in As the wave properties have been determined for each of the
part, to the unprecedented ∼1 s time cadence of the data set five equally spaced slits above the solar limb, we are able to
utilized. For comparison, previous investigations using data compare and study characteristics as a function of atmospheric
from the Swedish Solar Telescope provided cadences on the height. The mean values for displacement amplitude and
order of 5 s, which would make it very difficult for the lowest velocity amplitude are shown in Figure 6, where both
oscillation periods (<10 s) identified here to be detected. parameters can be seen to increase with height. By fitting a
Across all five defined slits, over 16,600 spicular threads linear line of best fit through the corresponding data points (see
were identified, of which 15,959 (95.9%) exhibit at least one the dashed red lines in Figure 6), the displacement amplitude
complete wave cycle. Of these examples, 8568 (51.5%) threads increases at a rate of 14.6 ± 0.8 km Mm−1, and the velocity
exhibit a single wave, 5770 (34.7%) consist of two superposed amplitude at 3.33 ± 0.08 km s−1/Mm. The conservation of
waves, and 1621 (9.2%) have three (or more) superposed energy flux requires a reciprocal relation between density and
waves. These proportions are similar to those found in velocity amplitude (see, e.g., the discussions in Stein &
transverse oscillations of coronal plumes using Auto-NUWT Leibacher 1974; Ebadi et al. 2012; de la Cruz Rodríguez et al.
by Weberg et al. (2018). 2013; Khomenko & Collados 2015; Grant et al. 2018; Houston
Two examples of the identified waves are shown in Figure 5. et al. 2018, 2020; Riedl et al. 2021). Employing spectro-
These are chosen as they have radically different periods and polarimetric inversions of the Ca II spectral line, Kuridze et al.
displacement amplitudes, consisting of ≈138 s and ≈358 km, (2021) revealed evidence that the mass density of spicules
respectively, for the top panel, with ≈20 s and ≈79 km, decreases exponentially with height, requiring the velocity
respectively, for the lower panel. Both waves are observed for amplitude to similarly increase to conserve energy flux. Hence,
longer than one full period. The wave identified in the top panel a quadratic fit is presented in the lower panel of Figure 6 using

5
The Astrophysical Journal, 930:129 (11pp), 2022 May 10 Bate et al.

Importantly, calculation of the phase velocities of the waves

embedded within the spicules allows for the eventual
calculation of their energy fluxes. The distribution of phase
velocities for the 135 propagating waves identified traveling
between the heights of 6200 → 6850 km are shown in Figure 7,
where waves propagating in both the upward and downward
directions are identified. Any waves displaying zero phase lags
(i.e., providing infinite phase speeds in Equation (1)) were
classified as standing modes. Due to the relatively small
number of standing modes present in our data set, this type of
wave is not included in the histogram depicted in Figure 7. We
must highlight that the distribution of upwardly and down-
wardly propagating waves shown in Figure 7 appear to
originate from the same population, with an approximately
Gaussian distribution encompassing waves propagating both
upwards and downwards. However, when examining the
Figure 7. A histogram showing the calculated phase velocities of the 135 energy flux carried by these waves, it is important to examine
propagating waves identified between the heights of 6200 and 6850 km above
the solar limb. The solid red line denotes the cumulative probability function. A
the direction of energy propagation, which is determined by the
bin width of 20 km s−1 was used for the creation of this histogram. sign of their associated phase speed. Thus, distinctions are
made between upwardly and downwardly propagating waves
a blue dashed line to show the potential synergy between for the sake of further analysis, but it must be emphasized that
expected mass density and velocity amplitude. However, we there do not seem to be two distinct populations present in
note that this is presented only for completeness, since it is Figure 7.
difficult to infer the true nature of the relationship with only It should be noted that any plasma flows within the spicules
five data points. The average periods do not show a similar will affect the apparent phase velocities of the measured kink
trend with atmospheric height, instead ranging within the same oscillations. In the case of upflowing plasma, the apparent (i.e.,
interval of 48.8–57.0 s for the five heights considered. measured) phase velocities of the upwardly propagating waves
In order to measure the phase velocity of these waves, it was will be related to (vph + U), while the downwardly propagating
necessary to identify the same feature across different heights. waves will have apparent phase velocities equal to (−vph + U),
This was achieved by extracting individual wave properties where vph is the true phase velocity and U is the velocity of the
from a certain atmospheric height and searching through the upflow (Nakariakov & Roberts 1995). This is similar to
wave catalog for waves with similar properties identified at an observations put forward by Grant et al. (2015), who examined
adjacent height. The properties considered for this study were the bulk plasma upflow within a magnetic pore and the
the equilibrium x-position of the spicule (±5 pixels subsequent effect this had on the apparent wave speeds of
or ±330 km), the midpoint time (whether or not the next sausage mode oscillations. Strong upflows are typically
spicule feature lay between the start and end times of the wave associated with type II spicules. However, the spicules
being compared), duration of the oscillation (±50%), and the observed in this study are likely not best characterized by this
frequency (±10%). Based on these criteria, around 140 waves classification (De Pontieu et al. 2007a). This effectively means
were found to be suitably similar between each set of adjacent that the apparent phase velocities of the upwardly propagating
heights, providing large number statistics with a similar waves can be considered an upper limit to their true phase
velocities. Conversely, in the case of the downwardly
proportion (tracked waves in relation to total identified waves)
propagating waves, this can be considered as a lower limit.
to that documented by Jafarzadeh et al. (2017).
As the velocity of any possible upflows are not known, the
The phase difference between all sets of waves identified at
measured phase velocities have been used in all further
adjacent heights was calculated using Fourier phase lag
calculations, but it is important to note that this will result in
analysis. The cross-power spectrum was calculated using the
the calculated energy fluxes being upper/lower limits for the
representative Fourier spectra of the two waves found at
upwardly/downwardly propagating kink waves.
adjacent heights (Bendat & Piersol 2000). The real part of the
Across the four sets of adjacent heights, the occurrence rates
cross-power spectrum (co-spectrum) was used to verify that
of upwardly propagating, downwardly propagating, and
each original Fourier spectrum had a peak at the same
standing mode waves were found to be 45%, 49%, and 6%,
frequency. The phase of the cross-power spectrum was then
respectively. This is in contrast to the occurrence rates found by
computed at the same frequency to determine the phase lag
Okamoto & De Pontieu (2011) of 59%, 21%, and 20%,
between the two heights (Vaughan & Nowak 1997). Finding respectively. However, the spicules examined by Okamoto &
this phase lag, f (in degrees), allows for the calculation of the De Pontieu (2011) were observed within a coronal hole, so may
phase velocity, vph (in kilometers per second), have different properties from those examined here. Impor-
tantly, our present study highlights a more equal balance of
360d
vph = , (1 ) upward/downward propagation, with fewer examples categor-
f ized as standing modes. The lack of standing mode detections
may also be a consequence of the incredibly high spatial and
where d is the height difference between the two slits in temporal resolutions of the HARDcam data set, since phase
kilometers, and  is the period of the wave in seconds (Jess precision is drastically improved as a result of the sub–1 second
et al. 2012b; González Manrique et al. 2020). cadence.

6
The Astrophysical Journal, 930:129 (11pp), 2022 May 10 Bate et al.

Table 2
Mean Phase Velocities for Each Set of Adjacent Heights that are Defined in
Table 1

Height Upward Downward

Phase Velocity Phase Velocity
(km) (km s−1) (km s−1)
4890 → 5550 128 ± 23 75 ± 12
5550 → 6200 131 ± 23 82 ± 23
6200 → 6850 139 ± 25 101 ± 15
6850 → 7500 147 ± 23 128 ± 23

It might initially be assumed that a roughly equal balance of

upward/downward propagation should be expected, due to the
high reflection coefficient of the transition region (Hollweg
et al. 1982). Liu et al. (2014) also observed downwardly
propagating transverse waves within solar spicules and note
that low-frequency periods (of ≈100 s) are expected to reflect
strongly in the transition region (Suzuki & Inutsuka 2005).
However, Okamoto & De Pontieu (2011) argue that such
reflection would result in more standing modes being observed
due to the superposition of upwardly and downwardly
propagating waves. This would create an imbalance in
observations, with more upward than downward propagations
detected. This superposition is, however, heavily dependent on
the height of the reflecting boundary, the phase velocity of the
upward wave, the lifetime of the spicule, and the time for
which the wave persists. If there is insufficient time for the Figure 8. Mean values of phase velocity shown with height above the solar
reflected wave to interact with the upward wave, due to any limb. The values for upwardly and downwardly propagating waves are shown
combination of the aforementioned criteria, then wave super- in the top and bottom graphs respectively. Errors are calculated using
bootstrapping.
position (and hence a standing wave) will not be observed.
Although a full characterization of the driving mechanisms The upward phase velocities appear to increase with
behind the downwardly propagating waves, as well as the clear atmospheric height at a rate of approximately 10 ± 15
domination of these waves with a phase speed around zero (see km s−1/Mm. However, due to the size of the associated
Figure 7), is beyond the scope of the present work, these are uncertainties (see the error bars in the upper panel of Figure 8),
important questions to be investigated in future studies. it is difficult to unequivocally stipulate the precise relationship.
For each of the four sets of adjacent heights, waves identified A more pronounced trend is present in the downward phase
as upwardly propagating were segregated from their down- velocities (lower panel of Figure 8), which appear to decrease
wardly propagating counterparts. It was hence possible to (as the height sampled decreases) at a rate of approximately
calculate the rate of change of phase velocity as a function of 24 ± 11 km s−1/Mm, implying that the wave slows as it travels
atmospheric height for both the upwardly and downwardly down the spicule and encounters more dense layers of the
propagating waves. Average phase velocities for each set of lower solar atmosphere.
adjacent heights are shown in Table 2 and plotted in Figure 8, With the velocity amplitudes and phase velocities of the
where the upper panel corresponds to the average phase oscillations measured, it was possible to estimate the energy
velocity of the upwardly propagating waves, while the lower flux associated with both the upwardly and downwardly
panel depicts the average phase velocity of the downward propagating waves. In order to calculate the energy flux, a
model for the density of the spicules with height is required.
propagation. The uncertainties shown in Figure 8 have been
Kuridze et al. (2021) observed a limb spicule and derived a
calculated following the “bootstrap” methodologies described
model of its density using the Non-LTE Inversion Code using
by Efron (1979). Due to the combined presence of traditional
the Lorien Engine (NICOLE; Socas-Navarro et al. 2015)
(periods 50 s) and high-frequency (periods <50 s) spicule inversion code. The final density model takes the form,
oscillations, it is challenging to assign basic standard errors to
the derived phase velocities, especially since the equivalence r ( y) = r0 e( y- h 0) L, (2 )
(or lack thereof) of the driving mechanisms responsible for
these characteristics have yet to be observationally and/or where y is the height above the solar limb, ρ(y) is the spicule
theoretically verified. As such, we apply bootstrapping density as a function of height, h0 is the base height, ρ0 is the
techniques to better constrain the confidence intervals of data density at the base height, and Λ is the density scale height.
following non-normal or unknown distributions (similar to that Values for our energy flux calculations were taken directly
presented by Simpson & Mayer-Hasselwander 1986; Desmars from Kuridze et al. (2021), where ρ0 ≈ 6 × 10−7 kg m−3,
et al. 2009; Yao et al. 2017). h0 = 2000 km, and Λ = 1500 km.

7
The Astrophysical Journal, 930:129 (11pp), 2022 May 10 Bate et al.

energy flux,
1
F » f r i v 2vph. (4 )
2
Taking the upper limit of the spicule density filling factor as
5% (Morton et al. 2012) allowed the energy fluxes to be
calculated for each adjacent set of heights, which are displayed
individually for all propagating waves (top panel) alongside
upwardly (middle) and downwardly (bottom) propagating
waves in Figure 9. For all waves examined, it can clearly be
seen that there is a decrease in energy flux with height,
indicated using solid black data points in the upper panel of
Figure 9. A linear line of best fit is presented using a dashed
black line in the upper panel of Figure 9, with a gradient of
−12,600 W m−2/Mm. However, an exponential fit would
perhaps be more appropriate, since the main factor for the
energy flux decrease is expected to be density stratification,
which is typically represented by a decaying exponential profile
with height (e.g., Verth et al. 2011). Due to the relatively small
number of data points under consideration, a linear fit has been
chosen for simplicity. Regardless of the fitting function
employed, the important message is that the energy flux of
the propagating transverse waves clearly decreases with
atmospheric height, hinting at some sort of damping and/or
dissipation process.
It is important to consider the effect of using a filling factor
of 5%. This means that Equation (4) estimates the energy flux
under the assumption that the waves are omnipresent, i.e., does
not take into account the sporadic nature of the observed wave
motion. In addition, as the waves are not seen to exist in all
spicules, the actual filling factor, f, should be reduced to
account for this effect. Thus, the estimation based on
Equation (4) gives us the upper limit of the energy flux in
the waves. However, as the filling factor is a multiplicative
term, this only affects the magnitude of any energy flux
estimations. The trends in energy flux examined with respect to
Figure 9. Energy flux estimations as a function of atmospheric height for all height are independent of any adjustment to the filling factor.
propagating waves (upper panel), upwardly propagating waves (middle
panel), and downwardly propagating waves (lower panel). The total energy For example, using the relatively low filling factor of 0.6%
flux provided by short/long-period waves is shown in black, while the suggested by Beckers (1972) will simply lower all energy flux
energy fluxes for short- (<50 s) and long-period (>50 s) waves are shown in and rate of change of flux values by a linear factor of 0.12 when
red and blue, respectively. The energy fluxes provided by the full set of compared to those values calculated with a filling factor of 5%.
waves (including upwardly and downwardly propagating) and for all
upwardly propagating waves are depicted, using a linear line of best fit, as a
The values presented in the text and within Figure 9 utilize a
dashed black line in the upper and middle panels, with gradients equal to filling factor of 5%, unless stated otherwise, and should
−12,600 W m−2/Mm and −13,200 W m−2/Mm, respectively. therefore be taken as upper limits.
For all upwardly propagating waves, we observe the energy
The energy flux, F, from transverse waves in a multiple flux flux to decrease as a function of height at a rate of
tube system can be calculated as, −13,200 ± 6500 W m−2/Mm, which is indicated in the middle
panel of Figure 9 using a dashed black line derived from a
1 linear least-squares fit. For completeness, it is estimated that
F » f (r i + re) v 2vgr , (3 ) energy fluxes in the range of 103–104 W m−2 are required to
2
heat the chromosphere (Withbroe & Noyes 1977). Hence, the
where f is the density filling factor, ρi is the density inside the total energy flux, in addition to the measured rate of energy flux
flux tube filled in by the spicule, ρe is the density outside the decay with height, are on the same order as the total energy
input required to provide basal heating to the solar chromo-
spicule, v is the velocity amplitude, and vgr is the group speed
sphere. Even considering the relatively low filling factor of
(Van Doorsselaere et al. 2014). For propagating kink waves, 0.6%, as suggested by Beckers (1972), the rate of energy flux
the group velocity can be approximated by the phase speed, decrease would be −1580 ± 780 W m−2/Mm, still within the
vph, as they are only weakly dispersive (Terradas et al. 2010; range that is needed to balance the radiative losses of the
Nakariakov et al. 2021). The internal density for spicules can chromosphere. By contrast, the energy flux for all of the waves
be assumed to be much larger than the external density, i.e., propagating downwards does not appear to depend on the
ρi ? ρe (Uchida 1961), providing a simplified equation for the height sampled (black data points in the lower panel of

8
The Astrophysical Journal, 930:129 (11pp), 2022 May 10 Bate et al.

Figure 9), with the energy flux estimates remaining consistent velocities and velocity amplitudes decrease with height (Soler
(∼ 4 × 104 W m−2) across the height range of approximately et al. 2011).
7500 → 4900 km above the solar limb. The similarity in the It has been proposed that, in order to supply the quasi-steady
rate of energy flux drop off in height is consistent between the effects needed to heat the solar atmosphere, the dissipation of
full set of waves (upper panel of Figure 9) and the upwardly short-period waves is of paramount importance (Hasan et al.
propagating ones (middle panel of Figure 9). This is to be 2005; Hasan & Van Ballegooijen 2008; Van Ballegooijen et al.
expected, since the downward energy flux remains approxi- 2011). The energy flux carried by both short-period (<50 s)
mately constant with atmospheric height. and long-period ( 50 s) waves between each set of adjacent
The decrease in upward energy flux with atmospheric height heights is shown in Figure 9 using red and blue data points,
may be due to at least three different factors: (1) physical respectively. In order to calculate the associated energy flux for
thermalization of wave energy into localized heat via the propagating wave modes, new filling factors were
dissipation mechanisms (e.g., Hollweg 1986; He et al. 2009; calculated by combining the previously used spicule density
Antolin et al. 2015, 2018; Okamoto et al. 2015, to name but a filling factor (5%; Morton et al. 2012) with the fraction of
few examples), (2) damping of detectable transverse waves waves that were found to fall into each relevant category (i.e.,
through the process of mode conversion, where kink mode <50 s or 50 s). The new filling factors were approximately
amplitudes decay as a result of the transfer of energy from 2.5%, which is a result of the 50 s boundary being very close to
transverse kink oscillations to azimuthal Alfvén motions the average period found at each height (see Table 1).
(Pascoe et al. 2010, 2012, 2013), and/or (3) reflection of the It can be seen from Figure 9 that the energy flux of the short-
waves downward at varying heights above the solar limb period waves is greater than that of the long-period waves for
(Hollweg et al. 1982; Suzuki & Inutsuka 2005). Tentative the full set of propagating waves (upper panel), and both the
observational evidence has shown that torsional Alfvén and upwardly propagating (middle panel) and downwardly propa-
kink waves may exist concurrently in spicules, providing gating (lower panel) waves. For the full set of propagating
credence for the applicability of mode conversion processes waves and the upwardly propagating waves, both the short- and
(De Pontieu et al. 2012). Previous modeling work by Sterling long-period waves show a similar energy flux decrease with
& Hollweg (1984) has shown that Alfvén waves within height as that for the total energy flux values. The energy flux
spicules can produce high-frequency signatures, including of both the short- and long-period downwardly propagating
periodicities of 112, 37, and 22 s for the fundamental, first, waves show a similar lack of dependence on atmospheric
and second harmonic resonant periods, respectively, which are height, which is consistent with the total energy flux
measurements. This suggests that both short- and long-period
similar to the periodicities found in our current work.
upwardly propagating waves have the potential to heat the solar
Employing simultaneous plane-of-sky imaging and line-of-
atmosphere, although the short-period waves have a larger
sight Doppler measurements will allow more precise defini-
energy flux across all heights, giving them a greater potential
tions of the embedded spicule wave modes, which will allow
capacity for heating.
the high-frequency Alfvén modes to be examined and
compared to the models put forward by Sterling &
Hollweg (1984). 4. Conclusions
In order to establish whether the wave energy is dissipated The results presented here represent a sizable increase in the
in the form of localized heating, measurements of thermal statistical population of examined transverse spicule oscilla-
processes in the vicinity of these spicules are necessary. This tions. Our use of data with a time cadence of ∼1 s also allowed
may be achieved using differential emission measures of for the identification of high-frequency waves, similar to those
optically thin coronal EUV observations directly above found by Okamoto & De Pontieu (2011), with periods as short
the spicules (McIntosh 2012; Vanninathan et al. 2012). as 10–20 s, only now with a significant increase in the
An alternative approach would be to use the Atacama examined population size. Observations with even higher
Large Millimeter/Submillimeter Array (ALMA; Wootten & spatial and temporal resolutions may allow for the detection of
Thompson 2009; Wedemeyer et al. 2016) to find the even shorter period and smaller-scale oscillations, and further
temperature of the spicules and the surrounding plasma extend the statistical distributions (see, e.g., Figure 4) down to
(Chintzoglou et al. 2021; Jafarzadeh et al. 2021; Henriques even smaller values.
et al. 2022). Importantly, the timing information related to the We examined the wave properties of spicule oscillations
decay of the spicule oscillations would need to be harnessed across multiple atmospheric heights, which facilitated the
to provide both spatial and temporal information to examine calculation of associated phase speeds, hence allowing us to
localized temperature fluctuations that may be a result of categorize the waves as either being upwardly/downwardly
thermalization mechanisms. While this is beyond the scope of propagating or standing. Almost an equal balance was found
the present work, it will form the basis of a follow-up study between upwardly (45%) and downwardly (49%) propagating
over the coming months. waves, in contrast to the earlier study by Okamoto & De
The downwardly propagating waves maintain an approxi- Pontieu (2011), who found that upwardly propagating waves
mately constant energy flux through a reduction in both were dominant in their time series. However, the observations
velocity amplitude and phase velocity as they travel down the presented here are in close proximity to the solar active region
spicule, visible in Figures 6 and 8, respectively. It is likely that NOAA AR12391 and may therefore have distinctly different
this is due to the wave interacting with the denser plasma at properties from the coronal hole observations examined by
lower heights above the solar limb, resulting in a slower Alfvén Okamoto & De Pontieu (2011).
speed in these regions (Okamoto & De Pontieu 2011). This is Directional information for the spicule waves allowed the
not unexpected, as the theoretical modeling of propagating kink calculation of their associated energy flux as a function of
waves in longitudinally stratified waveguides found that phase upwardly and downwardly propagating waves across a number

9
The Astrophysical Journal, 930:129 (11pp), 2022 May 10 Bate et al.

of atmospheric heights. Energy flux estimates are relatively References

consistent across all heights for the waves propagating in a
downwards direction. However, for the upwardly propagating Alfvén, H., & Lindblad, B. 1947, MNRAS, 107, 211
Alfvén, H. 1942, Natur, 150, 405
waves, a negative correlation with height is demonstrated, with Antolin, P., Okamoto, T. J., De Pontieu, B., et al. 2015, ApJ, 809, 72
the overall energy flux decreasing at a rate of −13,200 ± 6500 Antolin, P., Schmit, D., Pereira, T. M. D., De Pontieu, B., & De Moortel, I.
W m−2/Mm calculated with a spicule filling factor of 5% (or at 2018, ApJ, 856, 44
a rate of −1580 ± 780 W m−2/Mm using a lower-limit filling Beckers, J. M. 1972, ARA&A, 10, 73
Bendat, J. S., & Piersol, A. G. 2000, MeScT, 11, 1825
factor of 0.6%). The mechanism responsible may either be due Chintzoglou, G., De Pontieu, B., Martínez-Sykora, J., et al. 2021, ApJ, 906, 82
to thermalization of the mechanical wave energy or mode de la Cruz Rodríguez, J., Rouppe van der Voort, L., Socas-Navarro, H., &
coupling, although investigation of the proportional contribu- van Noort, M. 2013, A&A, 556, A115
tions of each mechanism are beyond the scope of this study. If De Pontieu, B., Carlsson, M., Rouppe Van Der Voort, L. H. M., et al. 2012,
even a small fraction of the wave energy carried in the ApJL, 752, L12
De Pontieu, B., Mcintosh, S., Hansteen, V. H., et al. 2007a, PASJ, 59, S655
transverse waves of the spicules examined is deposited as De Pontieu, B., McIntosh, S. W., Carlsson, M., et al. 2007b, Sci, 318, 1574
thermal energy, then it may significantly contribute to the Desmars, J., Arlot, S., Arlot, J. E., Lainey, V., & Vienne, A. 2009, A&A,
103–104 W m−2 requirements needed to balance the radiative 499, 321
losses of the chromosphere (Withbroe & Noyes 1977). Dunn, R. B. 1969, S&T, 38, 368
Ebadi, H., Zaqarashvili, T. V., & Zhelyazkov, I. 2012, Ap&SS, 337, 33
Edwin, P. M., & Roberts, B. 1983, SoPh, 88, 179
W.B., D.B.J., and F.P.K. acknowledge support from the Efron, B. 1979, AnSta, 7, 1
Leverhulme Trust via the Research Project grant RPG-2019- Erdélyi, R., & Fedun, V. 2007, Sci, 318, 1572
371. D.B.J. and S.D.T.G. wish to thank Invest NI and Randox González Manrique, S. J., Noda, C. Q., Kuckein, C., Cobo, B. R., &
Laboratories Ltd. for the award of a Research and Development Carlsson, M. 2020, A&A, 634, A19
Grant, S. D. T., Jess, D. B., Moreels, M. G., et al. 2015, ApJ, 806, 132
grant (059RDEN-1), in addition to the UK Science and Grant, S. D. T., Jess, D. B., Zaqarashvili, T. V., et al. 2018, NatPh, 14, 480
Technology Facilities Council (STFC) for the consolidated Hasan, S. S., & Van Ballegooijen, A. A. 2008, ApJ, 680, 1542
grant ST/T00021X/1. D.B.J. also acknowledges funding from Hasan, S. S., Van Ballegooijen, A. A., Kalkofen, W., & Steiner, O. 2005, ApJ,
the UK Space Agency via the National Space Technology 631, 1270
Programme (grant SSc-009). V.M.N. acknowledges support He, J. S., Tu, C. Y., Marsch, E., et al. 2009, A&A, 497, 525
Henriques, V. M. J., Jafarzadeh, S., Guevara Gómez, J. C., et al. 2022, A&A,
from STFC grant ST/T000252/1 and Russian Foundation for 659, A31
Basic Research (RFBR) grant (grant No. 18-29-21016). S.J. Hollweg, J. V. 1986, JGR, 91, 4111
acknowledges support from the European Research Council Hollweg, J. V., Jackson, S., & Galloway, D. 1982, SoPh, 75, 35
under the European Union Horizon 2020 research and Houston, S. J., Jess, D. B., Asensio Ramos, A., et al. 2018, ApJ, 860, 28
Houston, S. J., Jess, D. B., Keppens, R., et al. 2020, ApJ, 892, 49
innovation program (grant agreement No. 682462) and from Jafarzadeh, S., Solanki, S. K., Gafeira, R., et al. 2017, ApJS, 229, 9
the Research Council of Norway through its Centres of Jafarzadeh, S., Wedemeyer, S., Fleck, B., et al. 2021, RSPTA, 379, 20200174
Excellence scheme (project No. 262622). The Dunn Solar Jess, D. B., De Moortel, I., Mathioudakis, M., et al. 2012a, ApJ, 757, 160
Telescope at Sacramento Peak/NM was operated by the Jess, D. B., Mathioudakis, M., Christian, D. J., et al. 2010a, SoPh, 261, 363
National Solar Observatory (NSO). NSO is operated by the Jess, D. B., Mathioudakis, M., Christian, D. J., Crockett, P. J., & Keenan, F. P.
2010b, ApJL, 719, L134
Association of Universities for Research in Astronomy Jess, D. B., Morton, R. J., Verth, G., et al. 2015, SSRv, 190, 103
(AURA), Inc., under cooperative agreement with the National Jess, D. B., Shelyag, S., Mathioudakis, M., et al. 2012b, ApJ, 746, 183
Science Foundation (NSF). The authors wish to acknowledge Jess, D. B., Snow, B., Fleck, B., Stangalini, M., & Jafarzadeh, S. 2021, NatAs,
scientific discussions with the Waves in the Lower Solar 5, 5
Jess, D. B., Snow, B., Houston, S. J., et al. 2020, NatAs, 4, 220
Atmosphere (WaLSA; www.WaLSA.team) team, which is Khomenko, E., & Collados, M. 2015, LRSP, 12, 6
supported by the Research Council of Norway (project No. Kuridze, D., Morton, R. J., Erdélyi, R., et al. 2012, ApJ, 750, 51
262622), and The Royal Society through the award of funding Kuridze, D., Socas-Navarro, H., Koza, J., & Oliver, R. 2021, ApJ, 908, 168
to host the Theo Murphy Discussion Meeting “High-resolution Liu, Z. X., He, J. S., & Yan, L. M. 2014, RAA, 14, 299
wave dynamics in the lower solar atmosphere” (grant Makita, M. 2003, PNAOJ, 7, 1
McIntosh, S. W. 2012, SSRv, 172, 69
Hooke18b/SCTM). Morgan, H., & Druckmüller, M. 2014, SoPh, 289, 2945
Facility: DST (HARDcam; Jess et al. 2012a). Morton, R. J., Verth, G., Fedun, V., Shelyag, S., & Erdélyi, R. 2013, ApJ,
Software: MGN (Morgan & Druckmüller 2014), Auto- 768, 17
NUWT (Morton et al. 2013; Weberg et al. 2018). Morton, R. J., Verth, G., Hillier, A., & Erdélyi, R. 2014, ApJ, 784, 29
Morton, R. J., Verth, G., Jess, D. B., et al. 2012, NatCo, 3, 1315
Nakariakov, V. M., Anfinogentov, S. A., Antolin, P., et al. 2021, SSRv,
ORCID iDs 217, 73
Nakariakov, V. M., & Roberts, B. 1995, SoPh, 159, 213
W. Bate https://orcid.org/0000-0001-9629-5250 Nikol’Skii, G. M., & Sazanov, A. A. 1967, SvA, 10, 744
D. B. Jess https://orcid.org/0000-0002-9155-8039 Nikolsky, G. M., & Platova, A. G. 1971, SoPh, 18, 403
Okamoto, T. J., Antolin, P., De Pontieu, B., et al. 2015, ApJ, 809, 71
V. M. Nakariakov https://orcid.org/0000-0001-6423-8286 Okamoto, T. J., & De Pontieu, B. 2011, ApJL, 736, L24
S. D. T. Grant https://orcid.org/0000-0001-5170-9747 Pasachoff, J. M., Noyes, R. W., & Beckers, J. M. 1968, SoPh, 5, 131
S. Jafarzadeh https://orcid.org/0000-0002-7711-5397 Pascoe, D. J., Hood, A. W., De Moortel, I., & Wright, A. N. 2012, A&A,
M. Stangalini https://orcid.org/0000-0002-5365-7546 539, A37
Pascoe, D. J., Hood, A. W., De Moortel, I., & Wright, A. N. 2013, A&A,
P. H. Keys https://orcid.org/0000-0001-8556-470X 551, A40
D. J. Christian https://orcid.org/0000-0003-1746-3020 Pascoe, D. J., Wright, A. N., & De Moortel, I. 2010, ApJ, 711, 990
F. P. Keenan https://orcid.org/0000-0001-5435-1170 Pereira, T. M., De Pontieu, B., & Carlsson, M. 2012, ApJ, 759, 18

10
The Astrophysical Journal, 930:129 (11pp), 2022 May 10 Bate et al.

Pereira, T. M. D., De Pontieu, B., & Carlsson, M. 2013, ApJ, 764, 69 Van Ballegooijen, A. A., Asgari-Targhi, M., Cranmer, S. R., & DeLuca, E. E.
Pesnell, W. D., Thompson, B. J., & Chamberlin, P. C. 2012, SoPh, 275, 3 2011, ApJ, 736, 3
Poynton, C. 2003, Digital Video and HD: Algorithms and Interfaces Van Doorsselaere, T., Gijsen, S. E., Andries, J., & Verth, G. 2014, ApJ,
(Burlington, MA: Morgan Kaufmann) 795, 18
Riedl, J. M., Gilchrist-Millar, C. A., Van Doorsselaere, T., Jess, D. B., & Van Doorsselaere, T., Nakariakov, V. M., & Verwichte, E. 2008, ApJL,
Grant, S. D. T. 2021, A&A, 648, A77 676, L73
Rimmele, T. R. 2004, Proc. SPIE, 5490, 34 Vanninathan, K., Madjarska, M. S., Scullion, E., & Doyle, J. G. 2012, SoPh,
Rimmele, T. R., & Marino, J. 2011, LRSP, 8, 2 280, 425
Schou, J., Scherrer, P. H., Bush, R. I., et al. 2012, SoPh, 275, 229 Vaughan, B. A., & Nowak, M. A. 1997, ApJ, 474, L43
Secchi, A. 1875, Le Soleil (Paris: Gauthier-Villars) Verth, G., Goossens, M., & He, J. S. 2011, ApJL, 733, L15
Sharma, R., Verth, G., & Erdélyi, R. 2018, ApJ, 853, 61 Verth, G., & Jess, D. B. 2016, in Low-Frequency Waves in Space Plasmas, ed.
Shetye, J., Verwichte, E., Stangalini, M., & Doyle, J. G. 2021, ApJ, 921, 30 A. Keiling et al. (Washington, DC: AGU), 431
Simpson, G., & Mayer-Hasselwander, H. 1986, A&A, 162, 340 Weberg, M. J., Morton, R. J., & McLaughlin, J. A. 2018, ApJ, 852, 57
Socas-Navarro, H., de la Cruz Rodríguez, J., Asensio Ramos, A., Wedemeyer, S., Bastian, T., Brajša, R., et al. 2016, SSRv, 200, 1
Trujillo Bueno, J., & Ruiz Cobo, B. 2015, A&A, 577, A7 Wedemeyer-Böhm, S., Steiner, O., Bruls, J., & Rammacher, W. 2007, in ASP
Soler, R., Terradas, J., Verth, G., & Goossens, M. 2011, ApJ, 736, 10 Conf. Ser. 368, The Physics of Chromospheric Plasmas, ed. P. Heinzel et al.
Stein, R. F., & Leibacher, J. 1974, ARA&A, 12, 407 (San Francisco, CA: ASP), 93
Sterling, A. 1998, in ESA Special Publication 421, Solar Jets and Coronal Withbroe, G. L., & Noyes, R. W. 1977, ARA&A, 15, 363
Plumes, ed. T.-D. Guyenne (Paris: European Space Agency), 35 Wöger, F., Lühe, O. V. D., & Reardon, K. 2008, A&A, 488, 375
Sterling, A. C. 2000, SoPh, 196, 79 Wootten, A., & Thompson, A. R. 2009, IEEEP, 97, 1463
Sterling, A. C., & Hollweg, J. V. 1984, ApJ, 285, 843 Yao, J. M., Manchester, R. N., & Wang, N. 2017, MNRAS, 468, 3289
Suzuki, T. K., & Inutsuka, S. 2005, ApJ, 632, L49 Zaqarashvili, T. V., & Erdélyi, R. 2009, SSRv, 149, 355
Terradas, J., Goossens, M., & Verth, G. 2010, A&A, 524, A23 Zaqarashvili, T. V., & Skhirtladze, N. 2008, ApJL, 683, L91
Uchida, Y. 1961, PASJ, 13, 321 Zhang, Y. Z., Shibata, K., Wang, J. X., et al. 2012, ApJ, 750, 16

11