CHROMOSPHERIC RAPID BLUESHIFTED EXCURSIONS OBSERVED WITH IBIS AND THEIR ASSOCIATION WITH PHOTOSPHERIC MAGNETIC FIELD EVOLUTION

The line profiles of UFEs found by Lee et al. (2000) show absorption in the blue wing only, whereas the red wing could display either little deviation (e.g., the red profile in Figure 3(m)) or obvious intensity enhancement (e.g., the blue profile in Figure 3(m)) when compared to the quiet-Sun profile. The latter case is equivalent to the whole line shifting toward the blue. Strictly speaking, RBEs conform to the former kind of UFEs as they are characterized by a Doppler shift that only appears in the blue wing (e.g., Langangen et al. 2008). Nevertheless, as there may not be a fundamental physical difference between the two kinds of UFEs (Lee et al. 2000), we do not distinguish explicitly between UFEs and RBEs in this study. Here we develop an algorithm for detecting this type of events based on the spectral characteristics of features particularly in the blue wing. Specifically, the automated procedure examines the spectral line profile at each pixel to evaluate whether it has strong absorption in the blue wing only. In practice, the criterion is considered met when compared to the reference profile (the averaged profile over the entire FOV), the additional absorption in the blue wing is greater than 18% (∼3δ; here δ is the relative standard deviation of the image intensity) and that in the red wing is less than 6% (∼1δ), at Δλ = −0.4 Å and 0.4 Å  respectively, from the Ca ii 8542 line center. At each instance all the qualified, connected pixels are clustered together to form RBE features. Since RBEs may exhibit motion, the overlapping area of an RBE feature in at least three consecutive spectral scans (∼18 s) is defined as the region of this particular RBE. Finally, RBE regions that incorporate fewer than six pixels are removed, and directly neighboring RBE regions are merged. In this fashion, within our 95'' diameter circular FOV we identified 98 Ca ii 8542 RBEs distributed over the 11 minute time period from 2011 October 21 16:08:02 UT to 16:19:31 UT. Since our method takes full advantage of the spectral information in each pixel, it provides an appropriate and effective RBE detection by avoiding false and/or missed identifications, which might be an issue when only Doppler images are used.

In Figure 1, we use red contours to outline the detected 98 RBEs, 88 of which are in the FOV of Hinode observations. The RBEs are seen to be located mainly around the boundary of the supergranular cells (depicted by the dashed lines in Figure 1(b)) with some residing inside the cells, as also observed by Lee et al. (2000). Based on a visual inspection, the shape of these disk-center quiet-Sun RBEs is mostly round or elongated, similar to the result of Langangen et al. (2008) for a disk-center region. It is, however, noticeable that in all the previous RBE studies, the target region, either in the quiet-Sun or coronal holes, had obvious magnetic concentrations; accordingly, RBEs were mainly elongated and found in the inner edge of the conspicuous chromospheric rosette structures (Langangen et al. 2008; Rouppe van der Voort et al. 2009; Sekse et al. 2012, 2013a, 2013b). In contrast, we are looking at a very quiet region as shown by the highly sensitive Hinode LOS magnetogram; no well-formed fibril or rosette structure is discernible in the Ca ii 8542 line center image (Figure 1(a)).

In order to further characterize the RBE occurrence, we create λ-t plots by stacking the averaged line profile I(λ) within an RBE region over time. These kind of λ-t plots can also be viewed as $\upsilon _\mathrm{Doppler}$-t plots, where the Doppler velocity $\upsilon _\mathrm{Doppler}$ is related to the corresponding wavelength λ_c following

$$\begin{equation} \upsilon _\mathrm{Doppler}=\frac{c}{\lambda _0}(\lambda _\mathrm{c}-\lambda _0), \end{equation} \tag{ 1 }$$

where c is the speed of light and λ₀ is the wavelength of the resting line center (i.e., 8542.1 Å). For a more precise estimation of $\upsilon _\mathrm{Doppler}$ of the blueshifted spectral component, we follow the method introduced by Rouppe van der Voort et al. (2009):

$$\begin{equation} \upsilon _\mathrm{Doppler}=\frac{c}{\lambda _0}\frac{\int _{\lambda _\mathrm{min}}^{\lambda _0} (\lambda -\lambda _0)(I_\mathrm{ref}-I) d\lambda }{\int _{\lambda _\mathrm{min}}^{\lambda _0} (I_\mathrm{ref}-I) d\lambda } , \end{equation} \tag{ 2 }$$

where I is the intensity of the blueshifted profile and I_ref is the intensity of the reference profile. As the intensity fluctuations in the far blue wing can introduce large errors in estimating $\upsilon _\mathrm{Doppler}$, the integration is made from λ_min = 8541.0 Å to λ₀ = 8542.1 Å, which correspond to a $\upsilon _\mathrm{Doppler}$ ranging from −40 km s⁻¹ to 0 km s⁻¹. In addition, the integration is made over the absorption portion of the spectrum (i.e., where I < I_ref). The temporal evolution of spectra of two sample RBE events 1 and 2 (denoted in Figure 1) are drawn in Figures 2(l) and 3(l), respectively, in which λ_c corresponding to the derived $\upsilon _\mathrm{Doppler}$ at each instance of time is marked by the plus sign, with the time of the maximum $\upsilon _\mathrm{Doppler}$ colored blue. These plus signs clearly portray that a typical RBE event can be visualized in the temporal evolution of spectra by an asymmetric V-shaped excursion in wavelength toward the blue wing. In Figures 2(m) and 3(m), we also plot the blueshifted profiles, the reference profile (dashed black line), and their differences (most right group of lines) at some selected time instances (with the same color codes as those of the plus signs in Figures 2(l) and 3(l)), in both the scales of Δλ and $\upsilon _\mathrm{Doppler}$. These clearly demonstrate the characteristic, exclusive blueshifts of RBEs and their temporal evolution. We point out that around the time of maximum $\upsilon _\mathrm{Doppler}$, the average spectra of our defined RBE regions usually show a whole shift toward the blue, possibly indicating bulk motion. At other times, spectra with only an additional blue-wing component can be observed (see, e.g., the red profile in Figure 3(m)).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The detailed temporal evolution of spectra from IBIS enables us to investigate the temporal evolution of $\upsilon _\mathrm{Doppler}$ and acceleration/deceleration of RBEs in an intuitive and quantitative fashion. We find that the maximum $\upsilon _\mathrm{Doppler}$ of our 98 RBEs ranges from 12.4 to 22.3 km s⁻¹ with a mean of 16.1 km s⁻¹, and that the lifetime of RBEs in Ca ii 8542 Å ranges from 17.6 to 99.4 s with a mean of 39 s. These results agree with previous studies using CRISP data (e.g., Rouppe van der Voort et al. 2009; Sekse et al. 2012, 2013a), despite that the velocity is at the lower end. This could be because the spatial resolution of IBIS is lower than CRISP (Pereira et al. 2013), and because our region of interest is very quiet, much different from the regions studied before. In addition, the area of these RBEs ranges from 0.28 to 5.5 arcsec² (with a mean of 1.2 arcsec²) and is correlated with the RBE lifetime (with a correlation coefficient of 0.64), meaning that larger RBEs tend to live longer. Histograms of these physical properties are presented in Figure 4. Remarkably, we are also able to derive the properties of acceleration a_f and deceleration a_s of our RBEs in a straightforward way, by analyzing the slopes of edges of the V-shaped RBE feature in the $\upsilon _\mathrm{Doppler}$-t diagram. The study of tens of conspicuous cases gives an a_f of ∼200–700 m s⁻² with a mean of ∼400 m s⁻² and a following a_s ranging from ∼−90 to −240 m s⁻² with a mean of ∼−160 m s⁻², the latter of which is generally smaller than the gravitational acceleration of ∼−274 m s⁻² on the Sun. This suggests that our observed RBEs generally exhibit a fast acceleration followed by a slow deceleration toward blue. In some previous observations, evidence of acceleration of RBEs toward the end of their short appearance in chromospheric passbands was reported without a quantitative estimate (e.g., De Pontieu et al. 2007b; Rouppe van der Voort et al. 2009). Here we disclose, based on our high-cadence IBIS data, an additional deceleration phase that usually closely follows the initial acceleration. As a further comparison, the deceleration of jet-like dynamic fibrils in the chromosphere was measured earlier by fitting their paths in the distance–time plots (Hansteen et al. 2006; De Pontieu et al. 2007a), the result of which is similar to ours.

Zoom In Zoom Out Reset image size

It is worth mentioning that quite often the RBE signature seen in the temporal evolution of spectra seems to be related to the preceding and/or following redshift events, with a time gap of 1–2 minutes (see, e.g., Figures 2(l) and 3(l)). The overall shape appears sawtooth-like, reminiscent of the spectral signature of the consequence of acoustic shocks that propagate from the photosphere (e.g., Carlsson & Stein 1997; Vecchio et al. 2009). We cannot exclude the possibility that some of our detected RBEs, especially those located inside the cells, may include or be "modulated" by shocks. A quantitative analysis of these oscillations in relation to the RBE occurrence is not attempted in the present work. Nevertheless, many RBEs we found apparently occur at locations of magnetic structures as observed by Hinode (see further discussion below in Section 3.2), while previous studies suggested that shocks generally avoid magnetic elements (Vecchio et al. 2009). Alternatively, the sawtooth-like spectral structure can be due to the coexistence of RBEs and rapid redshifted excursions (see Figure 2 of Sekse et al. 2013b), the latter of which are observed on-disk owing to the torsional or transversal motions possibly harbored by spicules.

To shed light on the driving mechanism of RBEs, we examine the associated photospheric magnetic field evolution using the high-resolution LOS magnetograms from Hinode/NFI. As an illustration, the RBE events 1 and 2 as labeled in Figure 1 are presented as case studies. In Panels (a)–(i) of Figures 2 and 3, we show the time sequence of the underlying magnetic field, superimposed with the defined RBE region (red contour). The detected RBE features at selected instances are also overplotted in Panel (d). For an easier evaluation, we choose the related magnetic element of negative polarity and plot the corresponding temporal evolution of magnetic flux in Panel (k). We describe the events in detail as follows.

The RBE event 1 at the supergranule boundary is observed between about 16:17:04 UT and 16:17:38 UT. Right beneath the chromospheric RBE, a pair of magnetic elements p1-n1 of opposite polarity emerges from below the photosphere about seven minutes before the RBE occurrence (Figures 2(a)–(d)). The magnetic flux of the negative element n1 keeps increasing till 16:17 UT and then switches to a continuous decrease by ∼10¹⁷ Mx (Figure 2(k)), as n1 seemingly cancels with the adjacent positive field p1' (Figures 2(e)–(h)). The transition from flux emergence to cancellation is simultaneous with the RBE (Figure 2(k)), which migrates slightly southward (Figure 2(d)). The structure and evolution of these magnetic elements suggest that this RBE may undergo a miniature version of the eruption as in the classical coronal jet model (e.g., Shibata et al. 2007; Liu et al. 2011; Moore et al. 2011), where the emerging flux presses against the ambient opposite-polarity field to cause magnetic reconnection, driving RBEs. A consistent signature of energy release is the slight but discernible increase of the line core intensity cotemporal with the RBE occurrence (see Figure 2(m)). It is possible that the reconnection occurs at the low chromospheric level as the subsequent flux cancellation closely follows (see Section 3.3 for further discussion). The reduction of the p1' flux due to its cancellation with n1 is not obvious possibly due to the also emerging bipole p1'-n1'. It can be seen that n1' further cancels out with a southern positive element from ∼16:28 UT (Figures 2(g)–(j)), which might lead to another episode of RBE activity at roughly the same location (Sekse et al. 2013a).

Another RBE event 2 during about 16:15:59–16:16:40 UT is cospatial with the positive magnetic element p2 also located at the supergranule boundary (see Figure 1(b)). To the southwest lies the negative magnetic element n2 (Figures 3(a)–(b)), which is formed after coalescence of smaller magnetic patches (Lamb et al. 2013). Probably driven by convective flows, n2 moves toward and cancels with p2 starting from ∼16:12 UT (Figures 3(c)–(g)), until n2 dies out in 15 minutes with a total canceled flux of ∼2.5  ×  10¹⁷ Mx. Interestingly, the cancellation rate peaks around 16:16 UT, cotemporal with the onset of the RBE (see Figure 3(k)). The close spatiotemporal correlation is tempting to infer an intrinsic relation between the RBE and the flux cancellation. This kind of cancellation may be a consequence of magnetic interaction between network and intranetwork fields, as the supergranular convective flows sweep the latter toward the former (e.g., Wang et al. 1998; Gošć et al. 2014). Later, p2 cancels again with another negative element n2' (Figures 3(h)–(j)), which could be associated with a recurrent RBE event.

We attempt to explore whether the chromospheric RBEs are preferably associated with certain types of photospheric magnetic field evolution. To this end, we first resort to SWAMIS to automatically detect magnetic flux cancellation and emergence in our Hinode/NFI magnetic field data (Chen et al. 2015). We set the high threshold to 16.5 G (3σ) and the low threshold to 11 G (2σ), which SWAMIS uses to discriminate magnetic features based on the "clumping" algorithm. The recognized features are associated over time using a dual-maximum-overlap criterion, which can identify persistent features across image frames. For the best result, we added extra criteria that a feature must have an area of at least 4 pixels (041²) in each frame and its lifetime must be greater than ∼3 minutes (3 frames). The interaction among features is traced and classified into different types. A flux cancellation takes place when two adjacent (with a separation ⩽5 pixels, or 08) opposite-polarity features exhibit a flux decrease in at least 3 successive frames. The magnetic polarity inversion line between them is recorded as the cancellation site. A flux emergence is defined as a paired birth of opposite-polarity magnetic features, which are separated by at most 5 pixels. In the SWAMIS implementation, the requirement of flux conservation for the detected cancellation and emergence was set to be as permissive as physically reasonable—the changes in fluxes of the detected features only had to agree in sign, not in magnitude.

Next we define a circular region of radius r centered on each observed RBE, within which we search for the SWAMIS detected magnetic events during a time period τ. Higher limits of τ ≈ 1 hr and r ≈ 25 are used based on the following justifications. (1) Usually the small-scale magnetic reconnection occurs above the photosphere, and the newly formed inverse U-loop would retract and submerge below the photosphere at a descent speed of about 0.3–1 km s⁻¹ (Zwaan 1987; Harvey et al. 1999). This leads to a time delay between magnetic reconnection and the canceling magnetic features that ensued, which were found to be tens of minutes depending on the height of the reconnection region (e.g., Harvey et al. 1999; Yurchyshyn & Wang 2001; Chae et al. 2010). If magnetic loops submerge from an altitude of 1.2–1.5 Mm, which is the formation height of the Ca ii 8542 spectral line (Mein & Mein 1980), the time delay might be up to order of 1 hr. Here the implied possibility that RBEs could be driven by events of magnetic reconnection distributed over a range of heights was also suggested by Langangen et al. (2008). (2) Within 1 hr, a photospheric magnetic element can travel up to 25 given a velocity of ∼0.5 km s⁻¹ (Giannattasio et al. 2014).

As our 11 minute IBIS observation (16:08:02–16:19:31 UT) lies in the middle of the ∼2 hr Hinode/NFI data from 15:00–17:10 UT, we inspect magnetic cancellation events in the later half (16:08–17:08 UT) of the magnetogram sequence within the circular search region around each RBE. Similarly, flux emergence events are inspected in the early half (15:20–16:20) of the magnetic data. In Figures 5(a) and (b), we use cyan dots to mark the locations of the 378 cancellation events during 16:08–17:08 UT and those of the 246 emergence events during 15:20–16:20 UT as detected by SWAMIS, respectively. The 25 radius search region is also denoted by the yellow circles around the RBEs (red contours). As a result, we find that among the 88 RBEs in the Hinode FOV, 43 (49%) are associated with one or more cancellation; meanwhile, 34 (39%) are associated with one or more emergence. There are 20 (23%) RBEs that are associated with both flux emergence and cancellation, while 31 (35%) RBEs are associated with neither of them.

Zoom In Zoom Out Reset image size

To evaluate the significance of these correlations, we randomly select 88 locations in the Hinode FOV and compute the average number of magnetic events that can be found within a search region of radius r. This randomization is repeated for 100 times for a specific r and we also vary r up to 4''. The same calculation is also made for the RBE locations. In Figure 6, we plot r versus the average number of cancellation/emergence events that can be found, for both cases of the randomly selected locations (black lines with error bars) and the RBE sites (red lines). It appears that no matter what radius of the search region is used, the association of these RBEs with magnetic activities fall within two sigma of the random occurrence. As a further test, we repeat the evaluation after shortening the search time window τ to 40 and also 20 minutes. Still, the RBE-related magnetic cancellation (emergence) lies within two (three) sigma of the random occurrence. These results imply that the afore-described correlations may not be statistically significant; that is, these detected chromospheric RBEs are not statistically associated with photospheric magnetic events. Nonetheless, we caution that the lack of a statistical correlation might be partially attributed to the small data sample we analyzed (88 RBEs within 11 minutes), and also to the very quiet nature of this region. Obviously, more extended work needs to be done before definitive conclusions can be drawn.

Zoom In Zoom Out Reset image size