Документ взят из кэша поисковой машины. Адрес оригинального документа : http://star.arm.ac.uk/preprints/452.pdf Дата изменения: Mon Sep 12 13:10:07 2005 Дата индексирования: Mon Oct 1 21:53:57 2012 Кодировка: Поисковые слова: m 63

This is an unedited version of a paper accepted in A&A. To be cited as: A&A preprint doi:10.1051/0004-6361:20053714.

c ESO 2005

Very long period activity at the base of solar wind streams
M. D. Pop escu
1 2

1,2

, D. Banerjee3 , E. O'Shea1 , J. G. Doyle1 and L. D. Xia4

Armagh Observatory, College Hill, Armagh BT61 9DG, N. Ireland Astronomical Institute of the Romanian Academy, RO-75212 Bucharest 28, Romania 3 Indian Institute of Astrophysics, Koramangala, Bangalore 560034, India 4 School of Earth and Space Sciences, Univ. of Science and Technology of China, Hefei, Anhui 230026, China e-mail: mdp@arm.ac.uk; dipu@iiap.res.in; eos@arm.ac.uk; jgd@arm.ac.uk; xld@ustc.edu.cn Preprint online version: July 28, 2005 Abstract. Using time series data of sp ectral lines originating from a wide range of temp eratures in the solar transition region, ab ove a p olar coronal hole, from SUMER (Solar Ultraviolet Measurements of Emitted Radiation) on SoHO (Solar and Heliospheric Observatory), we rep ort on the detection of very long (170 min) p eriodic intensity fluctuations, ab ove the limb. Our data also reveal long p eriodicities (10-90 min), previously observed with other SoHO instruments. With the acoustic cut-off frequency implying a maximum allowable p eriod of 90 min, it is unclear whether these intensity fluctuations are due to waves or are the result of a recurrent magnetic reconnection process. Key words. Sun: p olar coronal holes ­ Sun: ultraviolet: SoHO ­ Sun: waves

1. Intro duction
It is known that quasi-perio dic fluctuations are present in open coronal structures. The high-cadence EIT/SoHO (171 °) observations of DeForest & Gurman (1998) repA resent, probably, the first detection of such MHD waves in polar plumes. These authors noticed bright features propagating outward at 75-150 km s-1 with perio ds of 1015 min. Based on the speed, which is close to the sound speed expected at T 1 в 106 K(cs = 147 km s-1 ), and on the density mo dulation inferred from the EUV brightness variation, they concluded that the observed fluctuations were compressive slow magneto-acoustic waves propagating along plumes. Banerjee et al. (2000a) investigated the temporal behaviour of polar plumes in the transition region (TR) line O v 629 ° with CDS/SoHO, noting the A presence of long period (20-30 min) compressional waves. Banerjee et al. (2001) extended this study to inter-plumes and reported on even longer perio dicities (70 min) over a wider temperature range. Ofman et al. (1997, 2000) searched for slow-mode compressional MHD waves with the UVCS/SoHO white light channel, further away from the Sun. They reported quasi-perio dic variations in the polarization brightness in both plume and inter-plume regions. Significant peaks were found around 1.6-2.5 mHz (610 min), with a coherence of 30 min. It is likely that the waves detected at 1.9 RSun by Ofman et al. (1997, 2000)
Send offprint requests to : Miruna D. Pop escu e-mail: mdp@arm.ac.uk or http://star.arm.ac.uk/preprints

with UVCS and the waves seen by DeForest & Gurman (1998) at 1.2 RSun with EIT are the same as the ones seen close to the limb by Banerjee et al. (2000a) using CDS. Teriaca et al. (2003) has presented evidence that the fast solar wind streams come from inter-plume regions. Here, we present the results of a search for the presence of perio dicities close to the solar limb, in an inter-plume region, from a time sequence of 14 h (probably the longest time series taken in a coronal hole with SUMER). This dataset has a wide temperature coverage of the TR, with lines originating from 8 в 104 to 6.3 в 105 K.

2. Data and metho d of analysis
The time sequence analysed in this Letter was taken in the South polar coronal hole with the SUMER spectrometer (Wilhelm et al. 1997) on 25 February 1997. The SUMER slit was fixed along the polar axis close to the pole. The slit is 1 wide (the solar-x direction) and 300 long (on solar-y ; North-South). The total time of the observation is 835 min (00:03:10.69 ­ 13:58:37.52), with 60 sec exposure time. In Fig. 1 we plot the SUMER pointing on an EIT image taken about half way through the observing time. SUMER had the rotational compensation switched off. The slit moved 47.6 at its base on the solar disk and 7.4 at its central position y = -950.25 (see horizontal lines in Fig. 1), and as we go towards the limb the movement of the Sun goes to zero. Therefore, in our region of interest ­ off-limb ­ the solar rotation is practically absent during the observation.

Original preprint version of an article accepted for publication in A&A


Pop escu et al.: Very long p eriod activity at the base of solar wind streams Fig. 1. The coronal hole from the Sun's South p ole in the light emitted by Fe xi i 195 ° obA served with EIT on 25 Feb. 1997. The vertical line represents the fixed p osition of the SUMER slit, with the horizontal lines showing how much the Sun rotated during the time series observation.

2

3. Results
Four examples of the wavelet transform applied to the flux of our spectral lines are illustrated in Fig. 2. In each example, the top panel shows the integrated flux variation in counts/60 s over time, with the overplotted line representing its trend, computed by a 50-pt running average. When applying the wavelet transform we did not de-trend the signal, as we are interested in the long-time perio dicities. The bottom left panel is the wavelet spectrum; and the bottom right panel is the global wavelet spectrum, which is the sum of the wavelet power over time at each oscillation perio d. The dark-coloured regions from the wavelet spectrum show the lo cations of the highest power. Crosshatched regions indicate the "cone of influence", where edge effects become important. The thick white line in the wavelet spectrum refers to the first maximum power. Only lo cations with a probability 95% are considered real (not due to noise). P1 is tested against the first maximum power found at any perio d in the randomized data. The top dashed line in the global wavelet spectrum marks the highest perio d cut-off ( 300 min). The multiple peaks in the global spectra indicate that the signal is composed of oscillations with different perio ds (see the horizontal lines in the right panel). These range from 11 to 66 min in the TR lines and from 18 to 93 min in Ne viii. We concentrate here on the very long (170 min) perio d. For all examples from Fig. 2, P1 170 min. We have chosen the flux at 5 above the continuum limb for the lower temperature lines because this very long perio dicity is more obviously visible (even by eye) than at other locations. For Ne viii, there is no such perio dicity at this location. Instead, a similar perio dicity begins to show further up, e.g. above the Ne viii limb (considering the continuum limb 0 , the Ne viii limb is at 11 and the N iv limb is at 3 ). The 170 min perio dicity in this line is best seen a few pixels above its own limb, e.g. at 15 . In summary, it is seen from 0 to 10-15 in the low temperature lines and from 10 up to 40 in Ne viii. The second to ol we have employed is the Fourier transform ­ see Fig. 3 that illustrates the results for the same variables as in Fig. 2. The peak at 0.1 mHz = 166.7 min is clearly seen. We do not take into account the smaller frequency (<0.1 mHz) as it would correspond to a third or a half length of our time series. As in the wavelet analysis, a range of shorter perio ds down to 11 min was observed. The significance of the peaks was tested using a 95% significance test, which in this case has a value of 18 (not plotted). The main peak at 170 min is statistically significant. The time resolution is high at low perio d values for both transforms (especially for the Fourier, whose resolution is 0.1 min). But it decreases towards higher values, e.g. the 170 min perio d has a resolution of ±9% in the wavelet and ±17% in the Fourier transform.

The image shows that the slit was positioned within an inter-plume lane, although close to a plume. For details on the data reduction pro cedures applied, see Xia et al. (2005) who used the same dataset for a morphological study of EUV spicules. We study the integrated flux variations of four lines: A A N iii 764.36 ° ( 8 в 104 K); N iv 765.15 ° (1.2 в 105 K); ° A O v 760.42 A (2.5 в 105 K); Ne viii 770.41 ° (6.3 в 105 K). Comparing the (background subtracted) flux with the intensity (computed as in Popescu et al. 2004), we find a 2-3% difference, with the flux being higher at maxima values. As regards the Doppler velo city, although at several off-limb lo cations it shows a clear blue/red structure (e.g. in macrospicules ­ see Xia et al. 2005), in fainter structures (spicules and especially in between them) the lines are not strong enough to show clear shifts. We search for perio dicities in the flux of the lines at different off-limb lo cations, averaging over 2 on solar-y . We produce power spectra using two different to ols: the wavelet and the Fourier transform. Details on the wavelet analysis, which provides information on the temporal dependence of a signal, are described in Torrence & Compo (1998)1 . For the convolution of the time series in the wavelet transform we cho ose the Morlet function as defined in Torrence & Compo (1998). To determine whether or not the oscillations found are real, we implemented the Linnell Nemec & Nemec (1985) randomization metho d that estimates the significance level of the main peak in the wavelet spectrum. The use of the randomization technique was done according to O'Shea et al. (2001). Here we use a randomization with 250 permutations. Secondly, we also use the Fourier transform (Jenkins & Watts 1968) in a similar way as Doyle et al. (1999). The significance levels for this transform were estimated using the fact that for a wide range of noise types (e.g. Poisson in this case) the noise power follows the 2 distribution with 2 degrees of freedom (Jenkins & Watts 1968). This allows one to construct significance levels using the integral probability of this 2 distribution, after first adopting the Leahy normalization, see Doyle et al. (1999) and references therein.
1

see http://paos.colorado.edu/research/wavelets/

Original preprint version of an article accepted for publication in A&A


Pop escu et al.: Very long p eriod activity at the base of solar wind streams

3

Fig. 2. Wavelet results for different lines, as lab eled. In each set, the top panel shows the variation of the line flux (averaged over 2 on solar-y ) over time, in counts/60s. The continuous overplotted line is a running average over 50 min. The wavelet p ower sp ectrum is given in the lower panel, with the white line showing the first maxima. P1 is the p eriod corresp onding to the first maximum p ower in the global wavelet p ower sp ectrum (right panel). In the global wavelet, the top dashed line is the maximum allowed p eriod (300 min). The horizontal lines indicate the p eriods that app ear in each dataset.

Fig. 3. The Fourier power spectrum for N iii 764 ° N iv 765 ° A; A; A O v 760 ° at 5 off-limb and Ne viii 770 ° at 15 off-limb. A

4. Discussion
Analyses of SUMER and UVCS data (Teriaca et al. 2003; Banerjee et al. 2000b) have shown that UV line widths are larger in inter-plumes than in plumes, suggesting interplumes to be the site from where the fast solar wind emanates. This argument is obviously supported by the fact that the plasma density is lower in inter-plumes. Thus, the perio dic intensity fluctuations detected here and by others in inter-plumes could play a relevant role in accelerating the fast streams of particles leaking from coronal holes, along open magnetic field lines. In addition to the long 10­90 min perio ds, the data presented here show a very long perio dicity of 170 min seen 5 above the limb in the flux of three TR lines (N iii 764 ° N iv 765 ° and O v 760 ° and 15 off-limb A, A A) in the Ne viii 770 ° line. This type of very long perio dicity A has not been reported before. Our observations were made off-limb in an inter-plume region. However, because of line of sight effects and the

proximity of plumes to the slit lo cation, we can not rule out contributions from both plumes and inter-plumes. Also, since we are dealing with single slit data, there is always the possibility that a feature is moving in and out of the slit. Judge et al. (2001) noted that over a 260 min interval, SUMER pointing was stable to within 0.2 thus well within the 1 slit width used here. Hence, if a feature was moving into and out of the field-of-view, it should be very intense in order to pro duce the required brightness changes. No such feature is observed; moreover, in Ne viii the intensity increase o ccurs over a 30 length along the slit, hence we rule out features moving in and out of the slit's field of view as an explanation for the intensity oscillations. Furthermore, all known SUMER corrections have being applied, including a correction for SUMER's thermo elastic oscillations (Rybak et al. 1999), which nevertheless affect only the wavelength shifts, therefore the Doppler velo city, whose variation we do not study here. As this is the first time we detect in a SUMER dataset the presence of long and very long perio d oscillations, we lo ok to other SoHO instruments to see if such variability has been observed. As mentioned in the Intro duction, various instruments have reported wave variability in open coronal structures, but not higher than 70 min. These perio ds have been generally interpreted as a signature of slow propagating (acoustic) waves. This is based on the fact that the slow mo des, being compressional in nature, pro duce a mo dulation in the density and EUV fluxes and thus, the observed integrated flux mo dulations (which is not the case for Alfvґ waves). Although longer-term varien ability of 8-27 hours (not necessarily waves) have been reported in filaments (Foullon et al. 2004), it is difficult to see the connection with the present observations.

Original preprint version of an article accepted for publication in A&A


Pop escu et al.: Very long p eriod activity at the base of solar wind streams

4

At first sight, the newly detected very long (170 min) perio ds might appear as just another mo de of the long (1090 min) oscillations, due to magneto-acoustic waves, but there are theoretical constraints to this idea (see discussion below). An alternative possibility is that they could be the result of a periodic pro cess, like for example some recurrent reconnection taking place in the chromospheric network boundaries. In an isothermal atmosphere, the acoustic cut-off frequency is given by ac = g /2cs (Roberts 2004). In the corona, we may take = 5/3, g = 0.274 km s-2 and cs = 200 km s-1 , which gives a cut-off perio d Pac (2/ac ) 91.7 min. Thus, under isothermal conditions, one would not expect to see any acoustic type of wave with periods as long as 170 min in the corona. But assuming the solar atmosphere as being purely isothermal is an approximation, thus the cut-off perio d calculated on such ideal conditions may not mimic the complicated temperature structure of the real atmosphere, and one should not rule out deviations from these numbers. For example, in a recent simulation, De Pontieu et al. (2005) showed that photospheric oscillations with acoustic cut-off frequency of 4.8 mHz or slightly lower can actually tunnel to higher up in the atmosphere as long as they are guided along an inclined magnetic flux tube. This nonverticality of the flux tube decreases the cut off frequency, allowing the lower frequency mo des to tunnel through. Polar plumes (inter-plumes) can very well act as wave guides. Thus, if slow waves with very long periods are generated somewhere in the middle atmosphere, this mechanism could allow them to eventually tunnel through to the corona. In a recent paper, Nakariakov et al. (2004) have demonstrated that some form of quasi-perio dic oscillations can yield a tadpole wavelet spectrum: a narrow spectrum tail precedes a broadband head (with some resemblance with our wavelet phase plots), but they pointed out that their quasi-perio dicity results from the geometrical dispersion of the guided fast magnetoacostic wave mo des (propagating within a closed magnetic lo op). In our case, if the oscillations are due to slow magneto-acoustic mo des (presumably they are propagating through open magnetic structures), the perio dicity could be prescribed by the source. For example, perhaps they are pro duced in the chromospheric network boundaries at the base of the coronal holes, where magnetic reconnections are usually present. Whatever the mechanism is that pro duces them, these 170 min periodicities add a significant new piece of information to the puzzle of mapping the perio dic phenomena of the solar atmosphere. Future mo dels need to explain the long perio d nature of the variability, lasting 550 min, with an amplitude of 15 to 20% of the total intensity.
Acknow ledgements. Armagh aided by the N. Ireland Dept. work was partially supp orted Irish Third Level Institutions Observatory's research is grantof Culture, Arts & Leisure. This by the Program for Research in for Grid-enabled Computational

Physics of Natural Phenomena (Cosmogrid) and by PPARC grant PPA/G/S/2002/00020. SUMER is financially supp orted by DLR, CNES, NASA and ESA PRODEX programme (Swiss contribution). SUMER and EIT are part of SoHO (ESA & NASA). We wish to thank the Royal Society, the British Council and DST-India for T&S supp ort within the UK and India. We also thank Ignacio Ugarte-Urra and Chia-Hsien Lin for help in using the wavelet analysis.

References
Banerjee, D., O'Shea, E., & Doyle, J. G. 2000a, Sol. Phys., 196, 63 Banerjee, D., O'Shea, E., Doyle, J. G., & Goossens, M. 2001, A&A, 377, 691 Banerjee, D., Teriaca, L., Doyle, J. G., & Lemaire, P. 2000b, Sol. Phys., 194, 43 De Pontieu, B., Erdґ i, R., & De Moortel, I. 2005, ApJ, 624, ely L61 DeForest, C. E. & Gurman, J. B. 1998, ApJ, 501, L217+ Doyle, J. G., van den Oord, G. H. J., O'Shea, E., & Banerjee, D. 1999, A&A, 347, 335 Foullon, C., Verwichte, E., & Nakariakov, V. M. 2004, A&A, 427, L5 Jenkins, G. & Watts, D. 1968, Sp ectral analysis and its applications (San Francisco: Holden-Day) Judge, P. G., Tarb ell, T. D., & Wilhelm, K. 2001, ApJ, 554, 424 Linnell Nemec, A. F. & Nemec, J. M. 1985, AJ, 90, 2317 Nakariakov, V. M., Arb er, T. D., Ault, C. E., et al. 2004, MNRAS, 349, 705 Ofman, L., Romoli, M., Poletto, G., Noci, G., & Kohl, J. L. 1997, ApJ, 491, L111 Ofman, L., Romoli, M., Poletto, G., Noci, G., & Kohl, J. L. 2000, ApJ, 529, 592 O'Shea, E., Banerjee, D., Doyle, J. G., Fleck, B., & Murtagh, F. 2001, A&A, 368, 1095 Pop escu, M. D., Doyle, J. G., & Xia, L. D. 2004, A&A, 421, 339 Rob erts, B. 2004, in ESA SP-547: SOHO 13 Rybak, J., Curdt, W., Ku cera, A., Schuehle, U., & WЁ l, H. Ё oeh 1999, in ESA SP-446: 8th SOHO Workshop, 579 Teriaca, L., Poletto, G., Romoli, M., & Biesecker, D. A. 2003, ApJ, 588, 566 Torrence, C. & Comp o, G. P. 1998, Bull. of the American Meteorological Soc., vol. 79, Issue 1, pp.61-78, 79, 61 Wilhelm, K., Lemaire, P., Curdt, W., et al. 1997, Sol. Phys., 170, 75 Xia, L. D., Pop escu, M. D., Doyle, J. G., & Giannikakis, J. 2005, A&A, in press

Original preprint version of an article accepted for publication in A&A