Wave amplitude modulation in fan loops as observed by AIA/SDO

Identification and characterization of magnetohydrodynamic (MHD) waves and oscillations are major thrust areas of research in the field of solar plasma-astrophysics. MHD waves are considered to be an important candidate for heating of the solar atmosphere (Erdélyi, 2004; Ofman, 2005; Taroyan & Erdélyi, 2009; Rajaguru et al., 2019; Liu et al., 2019). They may have significant roles in triggering of jets (De Pontieu et al., 2004; Chandra et al., 2015) as well as in the acceleration of fast solar wind (Cranmer, 2009).

One of the wave modes observed in the solar corona is slow mode magneto-acoustic waves. The propagation of this wave mode is largely confined to the magnetic field direction. They are compressional in nature, and propagate with speed that is less than the local sound speed in the solar corona. Such waves are often observed as fluctuations in intensity as well as in Doppler velocity (Wang et al., 2009; Gupta et al., 2012). These waves were first reported as propagating intensity disturbances (PIDs) by Ofman et al. (1997) in coronal plumes in off-limb observations recorded by the UltraViolet Coronagraph Spectrometer (UVCS, Kohl et al., 1995) on-board the Solar and Heliospheric Observatory (SOHO). Similar intensity oscillations in coronal fan loops were first detected by Berghmans & Clette (1999) using the Extreme-ultraviolet Imaging Telescope (EIT, Delaboudinière et al., 1995) on board SOHO in 195 Å channel, and later confirmed by De Moortel et al. (2000) in the observations recorded by Transition Region and Coronal Explorer (TRACE, Handy et al., 1999) in the 171 Å channel. Fan loops are magnetic plasma structures in the corona with temperature between 0.6–1 MK (Reale, 2010; Ghosh et al., 2017). They are found at the edges of active regions, and are mostly rooted in the umbra and umbra-penumbra boundaries of sunspots. They also remain anchored to the penumbra and non-sunspot (such as, plage) regions. From a statistical study, Kiddie et al. (2012) concluded that PIDs of sunspot origin are temperature dependent and have subsonic propagation speed.

Since the discovery of PIDs in coronal fan loops, a number of theoretical and observational studies have been performed to unravel their characteristics, propagation and dissipation and also their counterparts in the lower solar atmosphere (see, e.g., Nakariakov et al., 2000; De Moortel et al., 2002a, c; Roberts, 2006; De Moortel, 2009; Bogdan, 2000; Bogdan & Judge, 2006; Khomenko & Collados, 2015). Using wavelet analysis, De Moortel et al. (2002b) found that the periods of the PID are of the order of $282\pm 93$ s. However, careful observation of their footpoint locations showed that the coronal fan loops rooted in sunspot regions have a dominant period of $172\pm 32$ s, and the loops rooted in non-sunspot regions, such as plages are dominant in $321\pm 74$ s period oscillations. This distinction was explained in terms of the underlying driver at the photospheric level, exciting the loop footpoints.

In addition to the regular 3-min and 5-min oscillation, few studies have reported longer-period oscillations in the fan loops. Wang et al. (2009) has reported longer-period propagating waves at around 12 and 25 minutes along a fan-like coronal structure in the intensity (see also Marsh et al., 2009) and Doppler shift of the Fe XII 195  Å line using EUV Imaging Spectrometer (EIS, Culhane et al., 2007) on-board HINODE (Kosugi et al., 2007). Stenborg et al. (2011) showed the presence of waves in multiples of 8 minutes in coronal fan loops such as, in 16-min, 24-min, 40-min etc. It was speculated that the presence of the 8-min period could be attributed to beat phenomenon, which may occur due to the superposition of ubiquitous 3-min chromospheric and 5-min photospheric oscillations that leak into the corona through the inclined guide field that is provided by fan loops. However, they did not emphasise the excitation mechanisms of other longer-period oscillations that they detected in their observations.

In this paper, we perform a detailed analysis of intensity perturbations propagating along a fan loop system anchored within a sunspot. We track the evolution of these intensity oscillations by using the high spatially and temporally resolved data recorded by the Atmospheric Imaging Assembly (AIA, Lemen et al., 2012) on-board the Solar Dynamic Observatory (SDO, Pesnell et al., 2012). Our motivation is to understand how intensity oscillations with a range of periodicities interact, and affect the propagation characteristics of most dominant 3-min oscillations along fan loops. The rest of the paper is structured as follows. In Section 2, we describe the observations obtained for this study and briefly address the data reduction and preparation. We present the details of the different analyses acquired in this study, such as time-distance maps, Fouriér analysis, wavelet analysis and respective results in Section 3. Finally, we summarise our findings and conclude in Section 4.

The main aim of this work is to perform a multi-wavelength study of the dynamics of a fan loop system. For this purpose we have utilised the observations recorded by AIA using its UV channel, AIA 1600 Å and all the EUV channels, AIA 304 Å, 171 Å, 131 Å, 193 Å, 211 Å, 335 Å, and 94 Å. These channels are dominated with different spectral lines formed in a range of temperatures. For more information on the temperature sensitivity of different channels see O’Dwyer et al. (2010), Del Zanna et al. (2011), and Boerner et al. (2012). AIA provides full disk images with a cadence of 12 s in EUV and 24 s in UV. In addition, for contextual purpose, we use continuum data obtained using the Helioseismic and Magnetic Imager (HMI, Schou et al., 2012a, b), also on-board SDO. The AIA and HMI data are processed using standard processing software provided in the solar software (SSW; Freeland & Handy, 1998) distribution. The processed level 1.5 AIA and HMI data provides a pixel resolution of 0.6″.

The fan loop system, which is studied here, belongs to the active region NOAA AR 12553 observed on June 16, 2016. We have analysed the AIA observations recorded between 07:30:10 to 10:30:10 UT. We note that all the observations taken in different channels are rotated/de-rotated to the reference time at 9:15:00 UT for further analysis. Panel (a) in Fig. 1 displays a portion of the Sun’s disk image including AR 12553 observed in 171 Å passband. The yellow box encloses the fan loop system named as region of interest, a zoom-up of which is shown in panel (b).

In Fig. 2, we plot the regions of interest observed from HMI and AIA. Panels (a) and (b) show the HMI intensity and line-of-sight [LOS] magnetogram, respectively. The magnetogram is displayed within the magnetic field range of $\pm 1000$ G (with white indicating positive polarity and black indicating negative polarity). Panel (c) to (i) are taken by AIA in passband as labelled. The over-plotted cyan inner and yellow outer contours are plotted to visualise the umbra-penumbra and penumbra-plage boundaries, respectively. The umbra-penumbra contour is at the level of 38000 DN, whereas the penumbra-plage boundary is at the level of 53000 DN based on HMI continuum image. These contours are further over-plotted on the UV and EUV intensity images obtained in 1600 Å and 171 Å, shown in panels (c) and (e). Panel (e) reveals that some fan loops are rooted deep inside the umbra, and some are at the umbra-penumbra boundary.

Since our primary motivation is to study the wave dynamics of steady fan loops, we focus on the region that does not suffer from any flaring activity. With this condition in mind, we located three different regions on the fan loop system with white boxes, as shown in the panel (a) of Fig. 3 and studied the corresponding light curves over the following three-hour period. The resultant light curves are shown in panel (b) of Fig.3. The light curves are normalized with respect to the maximum intensities of each curve. The light curves conspicuously give the impression that the fan loop system does not show evolution more than 10% from the mean. Therefore, these could be considered for our study. The blue light curve taken from a location along fan loop 3 indicates some background intensity enhancement from time to time.

In order to study the existence and propagation of intensity disturbances along the fan loops, we employ time-distance analysis and construct artificial slits (slits 1, 2 and 3) along the three fan loops as shown in panel (a) of Fig. 3. We choose 171 Å image for the identification of fan loops as they are best seen at temperature of $\log\,[\textit{T/K}]=5.80$ (e.g., Brooks et al., 2011; Ghosh et al., 2017). We note that each slit is four pixels wide. We further emphasise that this analysis could be performed using one slit. However, to check the consistency of the results, we have used three different slits along different fan loops anchored in the umbra of the sunspot (see panel (a) in Fig. 3). We describe the results for slit 1 in the following, and provide the results for slits 2 and 3 in the Appendix for brevity.

In this work, we have studied the nature of PIDs observed in an active region fan loop system rooted in a sunspot umbra using AIA/SDO observations. We find that the PIDs are best observed in 171 Å data. The PIDs were also seen in other channels of AIA except for 335 Å and 94 Å. We have measured the phase speeds of the PIDs in all the AIA channels and found that the phase speeds are subsonic. This is suggestive that these are slow mode magneto-acoustic waves. We further found that the phase speeds measured in different channels are not dependent on temperature. This is in contradiction with the results obtained by, e.g., Banerjee et al. (2009); Krishna Prasad et al. (2012); Kiddie et al. (2012); Gupta (2014). A possible reason for this may be the broad temperature response of the AIA channels. Fan loops are structures of typically 1 MK; therefore it may be that the emissions from the lower-temperature components are contributing to these different AIA EUV emissions (Ghosh et al., 2017), thereby giving similar phase speeds.

Our result shows that the PIDs have longer detection length ($\approx 5.9-7.35$ Mm) in 171 Å and 193 Å, whereas a shorter detection length ($\approx 2.9$ Mm) in 131 Å and 211 Å. The similar detection length in 171 and 193 Å could be due to the similar temperature range covered by both these channels. We note that AIA 131 Å has a peak formation temperature closer to that of 171 Å in quiet conditions. Therefore, the shorter detection length observed in 131 Å may be due to the weaker signal-to-noise ratio (SNR) (¡ 10) than that of 171 Å (¿ 65). The shorter detection length observed in 211 Å channel may also be related to its SNR, which we find to be less than 20.

The light curves obtained at multiple locations along the loops show an increase and decrease in the amplitude of oscillations over time. Such variation is better seen in the de-trended light curves (see Fig. 6). Fouriér analysis on the original light curves reveals that there are several nearby frequencies within 5–8 mHz. The findings obtained here suggest that beat may be a possibility behind the observed amplitude modulation that are very likely created due to the interference of the nearby frequency components. We observe similar frequency peaks in all the AIA EUV coronal (except AIA 335 Å and 94 Å) as well as chromospheric channels. Lites (1986) has reported the various frequencies peaks identified in our study in chromospheric umbra, and Fludra (2001) reported them in transition region sunspot plumes. Centeno et al. (2006, 2009); Krishna Prasad et al. (2015) observed the presence of amplitude modulations in Fouríer filtered light curves and measured a modulation period of 20-27 minutes (Krishna Prasad et al., 2015). However, in our study, the amplitude modulations were prominent enough to detect them prior to any filtration (Fig. 6), which to our knowledge is the first such observational finding. Wavelet analysis also reveals an increase and decrease of the oscillation power around 3-min simultaneous to the observed amplitude modulation. We obtain the modulation period to be in the range 20-30 minutes.

Observation of similar amplitude variation at the lower solar atmosphere (right panels of Fig. 13) as observed by AIA 304 Å (log [T/K] = 4.7, Lemen et al. (2012)) supports one more possibility that may be the amplitude variability is caused by the underlying driver. Photospheric p-modes are, in general, assumed to be the cause of propagating slow magnetoacoustic waves observed in the corona (De Pontieu et al., 2005; Krishna Prasad et al., 2015; Zhao et al., 2016). Hence, the amplitude variability of the p-mode may affect and give rise to the particular amplitude variation of the propagating intensity waves observed along the fan loop. Observational studies have shown the effect oscillatory phenomenon in the lower solar atmosphere on the propagation of slow-magnetoacoustic waves along coronal fanloops (Jess et al., 2012; Sharma et al., 2017). However, further study is required both observationally and theoretically to understand the origin of such amplitude modulation in time series.

In addition to the amplitude modulation, we see presence of low period (high frequency) oscillations around 1.5-min (11 mHz) (see Fig. 14) in this study. We observe these high-frequency oscillations both at the chromospheric and coronal heights. Wang et al. (2018) first observed high-frequency oscillations by imaging observations at different heights above sunspot umbra in the range 10 – 14 mHz (around 1-min.). They related such high-frequency modes at coronal heights to perturbations at photospheric umbra. Krishna Prasad et al. (2017) reported frequencies around 13 mHz using IRIS 2796 Å in chromospheric sunspot umbra. We note that the observed high-frequency around 11 mHz is close to the double of the dominant (fundamental) frequency at around 5.56 mHz, and hence is similar to the second harmonics. However, one needs to be cautious to associate the high-frequency modes observed in our study to second harmonics, as harmonics are so far, observed in standing waves in coronal closed loops (Srivastava & Dwivedi, 2010). Verwichte et al. (2004) reported the detection of higher harmonics in coronal kink wave oscillations with a deviation in the frequency ratio of second to first harmonics from its canonical value of 2 in a homogeneous medium. Andries et al. (2005); Erdélyi & Verth (2007); Luna-Cardozo et al. (2012) proposed that such deviation may occur because of longitudinal density and or magnetic stratification along stratified expanding flux tubes like those in the corona, and therefore can be used as a seismological tool to measure the density scale height.