ABSTRACT
We examine the relationship between the high-frequency (425 MHz) type II radio burst and the associated white-light coronal mass ejection (CME) that occurred on 2011 February 13. The radio burst had a drift rate of 2.5 MHz s−1, indicating a relatively high shock speed. From SDO/AIA observations we find that a loop-like erupting front sweeps across high-density coronal loops near the start time of the burst (17:34:17 UT). The deduced distance of shock formation (0.06 Rs) from the flare center and speed of the shock (1100 km s−1) using the measured density from SDO/AIA observations are comparable to the height (0.05 Rs, from the solar surface) and speed (700 km s−1) of the CME leading edge observed by STEREO/EUVI. We conclude that the type II burst originates even in the low corona (<59 Mm or 0.08 Rs, above the solar surface) due to the fast CME shock passing through high-density loops.
Export citation and abstract BibTeX RIS
1. INTRODUCTION
Metric type II radio bursts are known radio signatures of coronal shocks. It is generally accepted that the type II burst is generated by a propagating shock that excites Langmuir waves, which converts into a radio wave at the local electron plasma frequency and its second harmonic (e.g., Melrose 1980). Since the ambient electron number density in the solar corona decreases with height, type II bursts appear as slow drifting lanes in the dynamic spectra from high to low frequencies. In most cases, the starting frequency of the type II burst is below 150 MHz (Maxwell & Thompson 1962) and its frequency drift rate is ∼0.3 MHz s−1 (e.g., Mann et al. 1995).
Since the first discovery of the metric type II burst by Payne-Scott et al. (1947), the debate on the origin of the type II burst has continued. One view is that the type II bursts are driven by fast coronal mass ejections (CMEs). Cliver et al. (1999) argued that a Morton wave in the chromosphere and a type II burst and EIT wave in the low corona have the same root cause in fast CMEs. To examine the possibility that all the type IIs are generated by only CME-related shocks, Cho et al. (2005) inspected 54 CME-type II events and concluded that at least 70%–80% of the metric type II bursts could be explained by a CME origin. For 41 type II bursts observed by the radioheliograph at the Gauribidanur observatory, Ramesh et al. (2012) compared the position of the bursts with the location of the associated CMEs and found that 92% (38/41) of the type IIs were located either at or above the CME leading edge. Gopalswamy et al. (2009) reported that the directly measurable height of the CME leading edge at type II onset time is similar to the height (∼0.5 Rs, from the solar surface) where the coronal Alfvén speed attains a minimum value (Gopalswamy et al. 2001). They concluded that the CME-driven fast-mode MHD shocks form very low in the corona and produce the type II bursts. In fact, direct observations of a CME-driven shock forming at a heliocentric distance of 1.2 Rs and the associated metric type II burst were reported recently (Gopalswamy et al. 2012a). The close relationship between type II bursts and CMEs was also established when limb type II bursts were compared with white-light CME observations (Gopalswamy et al. 2005). It was also found that the wavelength range of type II bursts is organized by the CME kinetic energy.
The opposite view is that the type II bursts are generated by flare-ignited shock waves (e.g., Harvey 1965). Magdalenić et al. (2008) present a metric type II burst on 1996 December 24, which was well timed with the associated flare, but the impulsive CME acceleration took place 10–20 minutes after the onset of the type II radio burst. The flare scenario looks favorable for the type II bursts with high starting frequencies (300–400 MHz in the fundamental band; Vršnak 2001). Statistical studies reported that most, if not all, high-frequency type II bursts are generated by flares since they occur in the low corona and their onset times often coincide with impulsive phases of flares (Cane & Reames 1988). For example, Vršnak et al. (1995) studied the relationship between onset times of the 28 high-frequency (>273 MHz) type II bursts and time characteristics of flares, and reported that the back-extrapolated launch time of the type II bursts coincides with the interval between the onset and the peak energy release of the flares. They concluded that the origin of the type II bursts is a blast wave ignited by the pressure pulse of a flare. This argument was supported by Shanmugaraju et al. (2009) who investigated the relationship between starting frequencies of high-frequency type II bursts (⩾100 MHz, 21 events) and the timing of flares (rising, duration, and delay) as well as the CME parameters (angular width, speed, and acceleration). They reported that there is a distinct relationship between the soft X-ray flare parameters and the starting frequencies, whereas the trend in the CME parameters shows low correlation. These conclusions are based on the onset of CMEs extrapolated back to the start time of the type II bursts assuming a constant speed of the CMEs, and the kinematics of type II shock estimated by adopting coronal density models. In fact, the real density distribution at the low corona can vary with time and location (Bemporad et al. 2003; Guhathakurta & Fisher 1998; Parenti et al. 2000), and we cannot ignore the fact that the observed CME kinematics might be different at the height where the type II bursts occur since the flare impulsive phase is often associated with CME acceleration phase (Zhang et al. 2001). In fact, this was illustrated using measurements available close to the Sun from the Solar-TErrestrial RElations Observatory (STEREO)/COR1, the Large Angle and Spectrometric Coronagraph (LASCO)/C1, and the Mauna Loa mark IV K-coronameter data (Gopalswamy et al. 2009, 2012b). In this respect, detailed investigations of the electron density distribution and the type II-associated-eruptive phenomena in low corona with high time and spatial resolutions are required to draw conclusions on the origin of high-frequency type II bursts.
In this paper, we present analysis of a high-frequency type II radio burst and address the question of whether the type II burst can be generated by a CME. The type II burst was detected by the Green Bank Solar Radio Burst Spectrometer (GBSRBS) on 2011 February 13. The starting frequency of the burst is unusually high (425 MHz in the fundamental lane).
First, we estimate the electron density along the path of eruption and match the density to the plasma frequency in the corona, which is the emission frequency of the type II burst. Previous studies on high-frequency type II bursts have usually adopted coronal density models to determine the height and speed of coronal shock. For example, Pohjolainen et al. (2008) used a barometric isothermal density model (Pohjolainen et al. 2007) to investigate the relationship between a high-frequency type II burst starting from 500 MHz in fundamental emission and a CME observed by Yohkoh/SXT, the Transition Region and Coronal Explorer, and Solar and Heliospheric Observatory/LASCO. They reported that the estimated speed of the type II shock was higher than the speeds of the white-light CME and EUV blob-like structure. As mentioned by Pohjolainen et al. (2008), the result of the shock kinematics derived from the type II emission is sensitive to the adopted value of the coronal density. Therefore, the real density information at the formation region of type II bursts, especially in high-frequency range, is crucial to investigate their relationship with flares and CMEs. To avoid the ambiguity introduced by using a fixed density model, we determine the real density distribution using six filter images taken by the Atmospheric Imaging Assembly (AIA; Lemen et al. 2012) on board the Solar Dynamic Observatory (SDO) and applying the automated emission measure analysis code recently developed by Aschwanden et al. (2011).
Second, thanks to the high time cadence observations from the Extreme Ultraviolet Imager (EUVI; Howard et al. 2008) on board STEREO, and SDO/AIA, the early phase of the CME during the high-frequency emission was well observed. It allows us to compare the kinematics of the type II shock and the CME directly in the low corona (<0.5 Rs, from the solar surface as in Gopalswamy et al. 2012a). In particular, the CME that appeared in STEREO Behind (STEREO-B) was devoid of projection effects because of its location (∼93° east of the Earth).
Third, we chose the event because of the clear fundamental–harmonic structure of the type II burst in the high-frequency range (300–1100 MHz). The main uncertainty in deriving the shock dynamics comes from the emission structure.
The plan of this paper is as follows: In Section 2, we briefly describe the event of 2011 February 13. In Section 3, we present the result of density measurement, location of the type II shock, and dynamics of the CME. Finally, Section 4 includes the discussion and conclusion.
2. OBSERVATIONS
The type II started in the early rise of the M6.6 flare that occurred on 2011 February 13 from AR 11158 located at the disk center (S20E04) seen from the Earth. The flare started at 17:28 UT, peaked at 17:38 UT, and ended at 17:47 UT as observed in GOES soft X-ray in the 1–8 Å channel.
Figure 1 shows the dynamic spectrum from GBSRBS, and the flux profiles of soft X-ray (black), hard X-ray (blue), and total radio (red) of GBSRBS. The frequency coverage of the spectrometer ranges from 5 MHz to 1100 MHz with 1 s time resolution (White 2007). As seen from the spectrum, the data from GBSRBS have a high signal-to-noise ratio with low background noise. The detailed description and data of GBSRBS are available at the Web site4 of the Green Bank observatory.
Figure 1. Dynamic spectrum of the type II radio burst followed by the type IV burst observed by GBSRBS from 17:34:17 UT to 17:57 UT on 2011 February 13. The white box in the upper panel denotes the high-frequency type II burst. Solid lines in the lower panel denote the GOES X-ray flux (black, 0.5–4 Å), RHESSI hard X-ray flux (blue, 20–50 keV), and total radio flux of GBSRBS (red, 200–1070 MHz), respectively. Fundamental and harmonic emission bands of the high-frequency type II burst are clearly seen during the rising phase of the soft X-ray flare. A group of emission features can be seen above 900 MHz. During the declining phase of the flare, type IV bursts followed and the shape of the type II burst becomes unclear due to the data gap between 50 MHz and 100 MHz by strong radio interferences. Our main interest is the high-frequency type II emission during the rising phase of the soft X-ray flare as marked with a white box.
Download figure:
Standard image High-resolution imageThe type II burst started from about 425 MHz at 17:34:17 UT and drifted to 10 MHz at 17:55 UT. The harmonic band starts from 850 MHz at 17:34:17 UT. The mean frequency drift rate of the type II during its initial stage (17:34:17–17:37 UT) as denoted by white box in Figure 1 is about 2.5 MHz s−1, much higher than the typical drift rate (∼0.3 MHz s−1) of type IIs reported by Mann et al. (1995). According to the drift rate dependence obtained in Gopalswamy et al. (2009), the drift rate at 425 MHz is expected to be around 0.9 MHz s−1. The drift rate of our type II burst is more than two times this value.
As seen in Figure 2, the early eruption of the CME was well observed by EUVI and COR1 on board the STEREO-B spacecraft. Since STEREO-A took only EUVI 195 Å images with a five-minute cadence, we use STEREO-B EUVI images at different wavelengths (171, 185, 284, 304 Å) to inspect the kinematic change of the CME with high time cadences. The CME in the STEREO-B image was devoid of projection effects because the eruption occurred near the disk center (S20E04) from Earth's view and STEREO-B was ∼93° east of the Earth. As viewed from STEREO-B, the source location of the CME is expected to be W89. The first appearance of the eruption was detected in the EUVI 304 Å image of STEREO-B at 17:34:19 UT. Considering the separation of STEREO-B from the Earth, we correct the observation time to the Earth time considering the position of STEREO-B. The corrected first appearance time of the CME is 17:33:59 UT, which is just before the start time (17:34:17 UT) of the type II burst. The first appearance height of the CME is about 0.05 Rs above the solar surface.
Figure 2. STEREO-B observations of the type II burst-associated CME (17:34–17:57 UT). The upper panels show the early phase of the CME observed by EUVI at 304 Å and 195 Å, and the lower panels show COR1 images with superposed EUVI 195 Å image.
Download figure:
Standard image High-resolution imageThe eruption can be traced by EUVI and the white-light observations from COR1. Figure 3 shows the height–time plot of the leading edge of the CME from 17:32:00 UT to 17:57:00 UT. The observation time is the corrected Earth time. The circles denote the measured heights and the plus symbol presents radial distance of the leading edge of the CME in three-dimensional space, which was calculated from twin STEREO/EUVI (195 Å) data using the solar soft routine called "scc_measure" developed by W. T. Thompson. Note that the measured heights from the STEREO-B observation are not so different from the radial distances of the CME in three-dimensional space from twin STEREO data. The CME shows slow rising (∼5 km s−2, 17:34 UT–17:35 UT), rapid acceleration (∼8 km s−2, 17:36 UT–17:37 UT), and constant speed (∼600 km s−1, 17:38 UT–17:57 UT). The initial rising and rapid acceleration phases coincide with the eruptive phase of the GOES soft X-ray flare and the low energy channel (6–12 keV) of the Reuven Ramaty High-Energy Solar Spectroscopic Imager (RHESSI; Lin et al. 2002) hard X-ray detector, which is consistent with the report by Zhang et al. (2001). It is noted that multiple peaks were detected in the 25–50 keV channel of the RHESSI hard X-ray detector during the slow rising and rapid acceleration of the CME as shown in Figure 3, which may be evidence supporting that the hard X-ray peaks were produced by the CME eruption (Cho et al. 2009).
Figure 3. Height–time measurements of the CME leading edge from STEREO-B observations, and time profiles of RHESSI hard X-ray in 6–12 keV (yellow line) and 25–50 keV (black line) channels. The dashed line denotes the GOES X-ray flux (0.5–4 Å). The CME heights are measured with respect to the solar surface at the position angle of 257° in which the fastest segments of the CME leading edge are shown. Filled color circles represent the CME heights measured at different wavelengths: 171 Å (green), 195 Å (blue), 284 Å (yellow), 304 Å (red), and white light (black). The plus symbols denote the radial distance of the CME leading edge in three-dimensional space calculated from twin STEREO data. The vertical red dashed line indicates the onset of the type II burst. We include an acceleration profile (blue dashed curve) to compare with the hard X-ray profile. The error bars on the acceleration (blue diamonds) are estimated by assuming a four pixel measurement error of the EUVI and COR1 data, respectively.
Download figure:
Standard image High-resolution image3. ANALYSIS AND RESULTS
3.1. Kinematics of Type II Shock from Density Models
Figure 4 is the detailed type II burst spectrum in the middle-frequency (100–300 MHz) and high-frequency (300–1100 MHz) ranges taken by GBSRBS from 17:34:17 UT to 17:37 UT. The spectrum shows simple structure of the burst with fundamental and harmonic bands. We can see the structures such as the splitting branch (Smerd et al. 1974) from the harmonic bands and emission patches above 900 MHz. It is not clear whether the patches are type III bursts related to the multiple peaks seen in the RHESSI hard X-ray channel, or the third harmonic of the type II burst, since no spectrum above 1100 MHz is available. In the harmonic band, three fragmented parts can be seen: before 17:34:52 UT, after 17:35:40 UT, and in between them. This kind of fragmented structure, which presents an inhomogeneity in the corona, was reported by Pohjolainen et al. (2008). Through the numerical MHD simulation of an erupting flux rope, they proposed that the fragmented emission may come from the shock passing through a system of dense loops overlying an active region.
Figure 4. Dynamic spectrum of the high-frequency type II burst between 17:34:17 and 17:37 UT on 2011 February 13 in the frequency range between 100 and 1070 MHz. Plus symbols denote the measured frequency and time of the type II emission at the harmonic and fundamental bands, respectively. The start time of the burst is marked with the vertical red dotted line. A band splitting (horizontal arrow) and fragments (vertical arrows) structures are seen in the harmonic band. Starting times and frequencies of the fragments ("F2" and "F3") are denoted by triangle signs. Faint emission patches above 900 MHz are likely the type III bursts or third harmonic of the type II burst.
Download figure:
Standard image High-resolution imageSince the harmonic lane shows more clear structure than the fundamental lane, we select the center frequencies of the harmonic lane at several times marked in Figure 4. The selected frequencies at the harmonic band are 850, 650, 450, and 350 MHz at 17:34:17 UT, 17:34:39 UT, 17:35:12 UT, and 17:35:36 UT, respectively. We divided the frequencies by two and marked them on the fundamental lane. It is found that they are quite well located on the fundamental lane as shown in Figure 4. We use the selected frequencies and times to estimate the formation heights of the type II shock and compare them with the kinematics of the CME.
The observed frequencies can be converted into heights and velocities of a shock if the distribution of electron density with height is known. Using density models by Newkirk (1961) and Saito (1970), we estimate the heights and speeds of the shock associated with the burst. The upper panel of Figure 5 shows the estimated shock heights from the emission frequencies marked with a plus symbol in Figure 4 by adopting onefold Newkirk (plus symbol) and twofold Newkirk (triangle symbol) models. The Alfvén speed profile in the bottom panel of Figure 5 is calculated by using the density models and magnetic field models of the quiet region (Mann et al. 1999) and the active region (Dulk & McLean 1978) of the Sun as was first done by Gopalswamy et al. (2001). As shown in Figure 5, it turns out that the shock is formed below the solar surface (unphysical) and accelerated with a speed greater than 1300 km s−1. The estimated speeds from the models are slower than the local Alfvén speed, which means that the speed is not fast enough to form a shock. These results strongly suggest that the analysis of the burst based on the density models does not give any reliable information on the kinematics of the type II shock.
Figure 5. Estimated formation height (upper panel) of the high-frequency type II burst and deduced shock speed (lower panel) using different coronal density models. In the upper panel, the heights of the shock are estimated using onefold Newkirk (plus) and twofold Newkirk (triangle) models. The solid, dotted, and dashed lines denote the electron density distribution from the Saito, onefold Newkirk, and twofold Newkirk models, respectively. Note that the shock forms below the solar surface (vertical dash-dotted line) when we apply the density models. In the lower panel, the speeds of the shock with respect to the solar surface are calculated using the onefold Newkirk (plus) and twofold Newkirk (triangle) models. The solid, dotted, and dashed curves denote Alfvén speed profiles estimated using the Saito, onefold Newkirk, and twofold Newkirk models with the magnetic field models of active (Dulk & McLean 1978) and quiet (Mann et al. 1999) regions of the Sun. The slow rising of the Alfvén speed above 0.8 Rs is due to our selection of the maximum value between the Alfvén speeds from the active region model and the quiet region model. Note that the estimated shock speeds from the density models are slower than the Alfvén speed.
Download figure:
Standard image High-resolution image3.2. CME Kinematics
The early stage of the CME eruption was well observed by SDO/AIA. Although the eruption can be seen at different AIA wavelengths, we select EUV 335 Å images because they present most clear eruption. Figure 6 shows the pre-event image (a) taken before the burst and pre-event-subtracted EUV 335 Å difference images ((b)–(d)) of the AR 11158 taken by SDO/AIA during the burst. According to Liu et al. (2012), the M6.6 flare is initiated at the center of AR 11158 where the mean horizontal magnetic field strength on the photosphere exhibited a significant increase and was produced through the tether-cutting reconnection process. They speculated that the reconnected large-scale field of the sigmoidal loops ("A" and "B" in Figure 6(a)) seems to erupt outward and becomes the CME. The eruption first appeared in the 335 Å difference image at 17:34:27 UT and propagated toward the southeast direction along the dotted line in Figure 6(a). The projected leading edge of the eruption on the disk is marked with the red cross in the figure. It appears likely that the shock itself is not resolved in the images. We measure the distance of the eruption along the line from the center of the flare (black plus) and find that its distance ranges from 48 to 85 Mm and the projected speed is about 440 km s−1. As seen in the 211 Å animation provided in the online journal we identified that there were dense loops. We cannot see clear traveling features across the northern loops during the type II burst time (17:34:17–17:37 UT) in 335 Å and the other AIA wavelengths.
Figure 6. SDO/AIA 335 Å observations before and during the high-frequency type II burst (17:34:17–17:36 UT). Panel (a) shows the SDO/AIA image taken before the eruption, and "A" and "B" denote the sigmoid loop associated with the eruption. Panels (b)–(d) show the difference images subtracting the background image (a) from the images taken during the type II burst. The fringe patterns seen in (b)–(d) are due to image saturation by the flare. The black plus symbol denotes the flare center and the red plus symbol represents the leading edge of the eruption. We measured the background electron density along the path between the flare center and the leading edge which is marked by a white dashed line in (a).(An animation of this figure is available in the online journal.)
Download figure:
Standard image High-resolution imageThe CME detected as a limb CME by STEREO observations is shown in Figure 7. There are loop systems in the north and south of the active region as denoted in Figure 7(a). It is noted that the CME appeared before the onset of the type II burst as reported in Gopalswamy et al. (2009, 2012a). The CME started to appear in the middle of the loop systems (Figure 7(b)) at 17:33:59 UT and crosses the southern loop during the type II burst time (Figures 7(c)–(f)). The tilted angle of the southern loop with respect to the solar surface is about 45° as shown in Figure 7(c). The leading edge of the CME can be seen until 17:42 UT in the field of view of STEREO-B/EUVI. It is likely that the bright feature in Figure 7(d) is the expanding flux rope that drives the shock. Due to the low spatial resolution of EUVI, we cannot resolve the shock structure itself. We measure the height and speed of the leading edge which is marked with a plus symbol in Figure 7. The height ranges from 34 to 104 Mm with the mean speed of 700 km s−1 during the burst time (17:34:17–17:36:00 UT). Because the type II burst occurred at higher frequencies, the above result indicates that the heights of the type II burst are even lower than in Gopalswamy et al. (2012a) who estimated the shock formation height (130–350 Mm or 0.2–0.5 Rs, from the solar surface) of the type II burst with a starting frequency of ∼300 MHz (harmonic component) by inspecting the AIA shock surrounding the CME flux near the Sun.
Figure 7. STEREO-B/EUVI image and difference images taken before ((a) and (b)) and during the high-frequency type II burst ((c)–(f)). The black plus symbol indicates the leading edges of the CME. It is noted that the CME appeared before the onset of the type II burst. During the type II burst, the southern loops are deflected and the CME disturbance is crossing the loops. The tilt angle of southern loops with respect to the solar surface is about 45°.
Download figure:
Standard image High-resolution image3.3. Kinematics of Type II Shock from Electron Density Measurements
The nose region of the CME is certainly favored for the emission location of the type II burst because of the high shock strength. Based on the above idea, we measure the electron density distribution along the propagating path of the CME's nose as denoted with black square symbols (path "A") in Figure 8(a) by using the numerical code developed by Aschwanden et al. (2011). In fact, the code is designed to generate emission measure and temperature maps automatically by using SDO/AIA image data sets in the six corona filters.
Figure 8. (a) Emission measure (EM) map determined by applying the automated emission measure code to SDO/AIA images in six corona filters taken before the eruption at 17:30 UT on 2011 February 13. The small boxes in (a) are the places where we measured the EM along three directions. Black, blue, and green dotted lines denote the directions of southern and northern corona loops, and the quiet region. Estimated electron densities along the paths are marked in (b) with plus symbols with different colors (black for path "A," blue for path "B," and green for path "C"). Dashed colored lines in (b) are the fits of the measurements (plus symbols) assuming a barometric density law for unmagnetized plasma, and the solid line indicates the Saito density model. Deduced starting distances of the radio fragments ("F2" and "F3") from the center of the flare are marked in (a) with arrows.
Download figure:
Standard image High-resolution imageUsing the code, we first carry out coalignment of AIA images in the six coronal EUV wavelengths by the limb fits and then calculate the Gaussian differential emission measure (DEM) distribution (see Equation (5) in Aschwanden et al. 2011) with best-fit values for the peak emission measure (EMp), peak temperature (Tp), and Gaussian temperature widths (σT) in each pixel. The DEM distribution can be reconstructed from the six filter fluxes (Fλ) in each pixel. A more detailed explanation on the procedure can be found in Aschwanden et al. (2011).
Figure 8(a) shows the peak emission measure of the AR 11158 just before the eruption. Since we have interest in the background electron density before the eruption, we select the AIA filter image data sets acquired around 17:30 UT (before the eruption) to avoid unwelcome artifacts by the diffraction pattern of the brightest flare kernels and pixel bleeding during flare peak time. Along the path (black square symbols) from the center (black plus symbol) of the flare region, we determine the density distribution on the path in the southern loops. For this, we select the mean values of EMp, Tp, and σT within the small boxes in Figure 8(a) and then determine the total emission measure (TEM) by the integral of the Gaussian DEM distribution over entire temperature range. To get an approximate value of the integral, we use the product of peak value (EMp) times the FWHM. The FWHM can be derived from the Gaussian temperature widths (σT) by multiplying by 2.35. The electron density distribution was estimated from the TEM by assuming that the line-of-sight distance of the observed region is the coronal pressure scale height (TEMl ≃ n2ll, where l is the line-of-sight length). For the scale height estimation, we use the hydrostatic isothermal scale height (λT = λ0(T/1 MK), with λ0 = 47 Mm). The estimated scale height is about 74,200 km for the mean temperature (1.58 MK) of the active region that is estimated using the code for the region of Figure 8(a). The estimated density along path "A" is denoted with black plus symbols in the right panel of Figure 8. The black dotted line in Figure 8(b) denotes the fits of density measurement by assuming a barometric density law for unmagnetized plasma. The estimated density ranges from 6× 109 to 1× 106 cm−3. As a comparison we also show, as a solid line in Figure 8(b), the model of Saito (1970). Note that the typical density models represent the height distribution of density, while our result gives the horizontal distribution of the density along the path of the CME's nose. Compared to the height distribution of density by the model of Saito (1970), the horizontal density from the center of the flare decreases much more steeply, and the density near the flare center is much higher than the density near surface of the Sun given by the model. As shown in Figure 8(a), there are similar loop structures in the north, which is also a potential source for the type II burst. We measure the density variation along paths "B" and "C," since the shock may be passing through the northern loops and the quiet region. The measured densities along these paths are marked in Figure 8(b) with blue and green plus symbols. We find that the density distribution along path "A" near the flare center (<0.2 Rs) is similar (or slightly higher) to that from path "B," but much higher than that from path "C."
We apply the distance distribution of the measured density to estimate the kinematics of the type II shock. Figure 9 shows the distances of the type II shock (upper panel) along three paths and the speeds of type II shock on the Alfvén speed profile (lower panel). We find that the type II shock forms at the distance of 41,760 km along path "A" corresponding to the distance between flare center and the leading edge of the first appeared SDO/AIA eruption, as shown in Figure 6(b). This distance corresponds to a height of 59,100 km if we consider the tilt angle (45°) of the southern loop structure. The estimated speed is about 1100 km s−1 (deprojected speed, about 1550 km s−1), which is slower than the speed of the shock estimated using typical density models, but faster than the Alfvén speed deduced by using our density measurement as shown in the lower panel of Figure 9. It is found that the estimated shock speeds from the density measurement along paths "A" and "B" are faster than the Alfvén speed, while its speed along path "C" is slower than the Alfvén speed and the shock formation distance along the path is unphysical.
Figure 9. Estimated shock distances and speeds using measured electron densities for three different paths. The upper panel shows the measured density fits (black for path "A," blue for path "B," and green for path "C") and the deduced shock distances from each plus symbol in Figure 8(a) by applying the density fits to type II emission frequencies. The lower panel shows the shock speeds (black, blue, and green diamonds for paths "A," "B," and "C," respectively) estimated using the density measurements. The dash-dotted color lines denote the Alfvén speed profiles calculated using the measured density fits (dash-dotted color lines in the upper panel) and the magnetic field models of active (Dulk & McLean 1978) and quiet (Mann et al. 1999) regions as was done early by Gopalswamy et al. (2009). Slow rising of the Alfvén speed above 0.4 Rs in the black dash-dotted line is due to our selection of the maximum value between the Alfvén speeds from the active region model and the quiet region model.
Download figure:
Standard image High-resolution imageFigure 10 shows height and distance comparisons among the leading edge of the STEREO-B/EUV CME, the nose of SDO/AIA eruption across the southern loops, and the fronts of the type II shock. Black and blue crosses denote the estimated shock distances along paths "A" and "B," respectively. We find a close relationship between the type II shock and the EUV CME: (1) they appeared at a similar time (17:34 UT); (2) the type II shock is located ahead of the CME nose from SDO/AIA and the CME leading edge from STEREO-B/EUVI; and (3) the distances among them are increasing while the CME propagates. From the density measurement along path "A," we find that the type II shock is located at the distance between 0.06 Rs and 0.20 Rs from the flare center. During the type II burst time (17:34:17 UT to 17:35:36 UT), the leading edge of the CME seen from EUVI ranges from 0.05 Rs to 0.16 Rs above the solar surface, and the CME's nose seen by SDO/AIA is located between 0.07 Rs and 0.12 Rs from the flare center. If we consider the tilt angle (∼45°) of the southern loops, the deprojected heights of the shock are estimated in a range between 0.08 Rs and 0.26 Rs above the solar surface. We cannot compare the distance type II shock front along path "B" to the kinematics of the SDO/AIA eruption along the path, because any significant traveling features cannot be seen along the path during the type II burst time (17:34:17–17:37 UT).
Figure 10. Height and distance comparisons among the leading edge of the STEREO-B/EUV CME, the nose of the SDO/AIA eruption, and the front of the type II shock. The leading edges of the STEREO-B/EUV CME (squares) are measured from the solar surface, and the projected noses of the SDO/EUV eruption along path "A" (filled circles) and the fronts of the type II shock are estimated from the density along paths "A" (black cross) and "B" (blue cross), respectively.
Download figure:
Standard image High-resolution image
The fragmentary structures in the type II emission are likely related to observed brightness (density) variations in the southern loops. As shown in the dynamic spectrum (Figure 4) of the type II burst, the second ("F2") and third ("F3") fragments start from 580 MHz at 17:34:52 UT and from 330 MHz at 17:35:42 UT, respectively. Applying the density distribution along path "A," we deduce starting distances of the second and third fragments from the center of the flare and find that they start from ∼0.1 Rs ("F2") and ∼0.2 Rs ("F3"), respectively. The deduced distances from the radio spectrum are similar to the spatial scales of the loop structures as shown in Figure 8(a).
4. DISCUSSION AND CONCLUSION
We have analyzed the high-frequency type II radio burst of 2011 February 13 to check if the high-frequency type II burst is generated by the associated CME. We found that applying typical coronal models is invalid to analyze the burst, but the formation of the burst in the low corona can be successfully explained by the CME-driven shock, if we use the measured coronal electron density. The early eruption of the CME at the start time of the burst was well observed by SDO/AIA and STEREO/EUVI. From the AIA observation, we found that the CME-driven shock propagates through the high-density loops during the burst time. The background density measured on the path of the CME propagation steeply decreases from high (∼6 × 109) to low (∼1 × 106 cm−3). It is found that the kinematics of the type II shock deduced from our density measurement are similar to those of the CME observed by STEREO/EUVI and SDO/AIA.
Surrounding the outward moving CME, the shock is formed as a dome-like structure at 17:34:17 UT (Figure 7). The type II burst is likely generated at the oblique region of the CME-driven shock where the shock sweeps across the multiple southern loops tilted 45° to the surface. Our density measurement shows that the density of the inner loops is much higher than the others, and the density rapidly decreases from the inner to the outer loops. The high electron density and magnetic field structure near the apex of the loops provide the shock with the low Alfvén speed to form the type II burst. It is likely that the fragment structures of the burst are associated with the density changes of the loops, and the start of the type II burst in the unusual high frequency may be associated with the high dense loops. This is consistent with the results of the numerical simulation by Pohjolainen et al. (2008) who argued that the fragment part of the type II burst can be formed when a CME-driven coronal shock passes through a system of dense loops overlying the active region.
Based on our density estimation and low coronal observations with high time and spatial resolutions, we conclude that unusual high-frequency type II bursts were generated in a special situation when the shock driven by a fast CME propagates into high-density loops in the low corona.
The primary result of this paper is that the type II burst can occur even in the inner corona (59 Mm or 0.08 Rs, above the solar surface) by the CME-driven shock. A more extended investigation is required to draw a more definite conclusion on the location of the type II burst emission by determining two-dimensional space and height distributions of the electron density in the low corona.
This research was supported by a NASA LWS TR&T program. National Radio Astronomy Observatory (NRAO) is operated for the NSF by associated universities, Inc., under a cooperative agreement. K.S.C. is indebted to M. Aschwanden for assistance with the coronal density estimation. We thank the referee for helpful comments, which improved the presentation of the paper. This work was partially supported by the "Development of Korea Space Weather Center" project of KASI and the KASI basic research fund.