Oscillations and running waves observed in sunspots

Transcription

1 Astron. Astrophys. 354, (2000) ASTRONOMY AND ASTROPHYSICS Oscillations and running waves observed in sunspots E.B. Christopoulou 1, A.A. Georgakilas 1, and S. Koutchmy 2 1 Thessalias 13, Petroupoli, Athens, Greece (algeorg@titan.astro.noa.gr; a304@sapfo.astro.noa.gr) 2 Institut d Astrophysique de Paris, CNRS, 98 bis boulevard Arago, Paris, France (koutchmy@iap.fr) Received 18 October 1999 / Accepted 1 December 1999 Abstract. In order to study umbral oscillations, running penumbral waves and the relationship between them, we analyzed CCD, high resolution, sunspot observations obtained at center and the wings of the Hα line and the Fe I 5576 Å line. The UBF filter was used in order to produce high cadence sequences of filtergrams. Images were processed to remove the sharp intensity gradient between the umbra and the penumbra. They show the waves to start out around umbral oscillating elements and to propagate outwards forming concentric cycles around the elements. The waves appear to propagate beyond the outer edge of the photospheric penumbra, in the superpenumbra, where they dilute. Comparing images in 9 wavelengths along the Hα profile we found out that the waves are definitely better observed near the Hα center and near the blue wing 0.35 Å. This indicates a possible vertical upward mass motion in the oscillating penumbral structure and that the oscillation is not symmetric about zero. We found different oscillating modes. Standing umbral oscillations are dominant in the umbra and inner penumbra; their frequency is around 6.5 mhz. Similar oscillations are observed in the penumbra - superpenumbra boundary but with considerably lower frequency (2 mhz). Oscillations are absent or have reduced magnitude in the central part of the penumbra. Penumbral waves are running waves propagating with a constant phase velocity around 13 km s 1 ; their frequency is remaining constant over the whole penumbra, in the band of 3 mhz. We produced time slice images which show, that there is not a clear relationship between umbral oscillations and running penumbral waves. Key words: Sun: sunspots Sun: oscillations 1. Introduction Sunspot umbrae show oscillatory behavior, identifiable as intensity and velocity variations in photospheric and chromospheric lines. There are observed two, apparently different, types of oscillations in sunspot umbrae, with periods around 5 minutes and 3 minutes respectively. Three minute umbral velocity and Send offprint requests to: E.B. Christopoulou intensity oscillations in photospheric and chromospheric lines, were first reported by Giovanelli (1972), Bhatnagar & Tanaka (1972), and Beckers & Schultz(1972). Oscillations within the 5-minute band were first reported by Bhatnagar et al. (1972), Shröter & Soltau (1976), and Soltau et al. (1976). According to Lites (1992), who summarized recent results the oscillatory motions measured in spectral lines formed in the lower photosphere of umbrae show a dominant power peak centered around 5-minute periods. In the chromosphere the dominant period of oscillations is in the 3 minute band; 5-minute umbral oscillations have either very low amplitudes or they are not present at all. Lites & Thomas (1985) found that the 3 minute umbral oscillation has a character of a coherent vertically standing wave in the photosphere and chromosphere. Kentischer & Mattig (1995) found indications that the 3 minute oscillations have their origin in photospheric layers, confirming the above results. It is generally believed that the 3 minute oscillations is a resonant mode of the sunspot magnetic flux tube either of chromospheric or photospheric resonance, while the 5-minute oscillations is the response of the sunspot to forcing by the p-mode oscillations in the surrounding convection zone (Thomas & Weiss, 1992; Chitre, 1992). Running penumbral (rp) waves were first reported by Zirin & Stein (1972) and by Giovanelli (1972). Zirin & Stein identified the rp waves as running intensity waves, observed in Hα center, with period around 300 sec and constant velocity around 10 km s 1. On the other hand, Giovanelli (1972) studied the waves using dopplergrams, obtained from the subtraction of the redwing of Hα from the blue one. He found that penumbral waves are due to a vertical velocity oscillation, symmetric about zero, and that they propagate outwards with a velocity of about 20 km s 1. Since then, the rp waves are observed in Hα as narrow alternate bright and dark bands, concentric to the edge of the umbra. The alternate darkening and brightening in the wave is probably produced mostly by Doppler shifting of the Hα absorption profile in and out of the filter bandpass due to periodic vertical mass motion (Moore 1981). Some authors refer to them using the more general term penumbral oscillations. The horizontal propagation velocity of penumbral waves is typically in the range km s 1. Zirin & Stein (1972) found a constant

2 306 E.B. Christopoulou et al.: Oscillations and running waves observed in sunspots Fig. 1. Hα image of a large isolated sunspot observed near disk center on August (top). The same image processed in order to enhance running penumbral waves (bottom) propagation velocity; on the other hand Giovanelli reported indications of an acceleration, that however he could not verify. Lites (1988) from observations in Ca II (8498 Å) line found a decrease of the frequency of oscillation with distance from the umbra. Brisken & Zirin(1997) found, recently, that the waves decelerate from 25 km s 1 near the inner part of the penumbra to 10 km s 1 near the outer edge of the penumbra. Finally Sigwarth & Mattig (1997) reported outward moving waves with higher phase velocities (10 to 40 km s 1 ) in the inner and with small phase velocities (2-3 km s 1 ) in the outer penumbra. Lites (1988) observed in Fe I 5434 Å Doppler amplitude images, a ring of low oscillatory amplitude halfway between the inner and outer edge of the penumbrae. He interpreted this ring as a result from the p-mode oscillations becoming increasingly more important in the outer penumbrae of sunspots. According to Zirin & Stein (1972) the rp waves are best observable in Hα center. Moore & Tang (1975) found that they are much more noticeable in the blue wing. Finally, according to Beckers rp waves are best observable in Hα center and in Hα 0.3 Å, they are almost absent in Hα+0.3 Å and hardly identifiable in Hα±0.5 Å (Discussion following Giovanelli s presentation in IAU symposium 156, 1974). The wavelength where they are best observed could clarify the mechanism that makes them identifiable and give us indications about their nature. Lites (1988) from Doppler movies of sunspots in Fe I 5434 observed a continuity of the disturbances across the umbrapenumbra boundaries. On the basis of the above observation he proposed that either the penumbral waves in the inner penumbra are driven by the umbral oscillations or the umbral and penumbral oscillations share a common physical basis. The above

3 E.B. Christopoulou et al.: Oscillations and running waves observed in sunspots 307 Fig. 2a i. Processed images showing the appearance of running waves in a Hα center, b Hα 0.35 Å, c Hα+0.35 Å, d Fe I (5576 Å), e Hα 0.5 Å, f Hα+0.5 Å, g Dopplergram in Fe I ±0.012 (5576 Å), h Hα 0.75 Å, i Hα+0.75 Å

4 308 E.B. Christopoulou et al.: Oscillations and running waves observed in sunspots Fig. 3. Subtracted images that show the waves to start out inside the umbra around oscillating elements (believed to be the origin of rp waves) and to propagate outwards Fig. 4. Subtracted images in Hα center, that show the propagation of rp waves, beyond the outer limit of the penumbra, in the superpenumbra. picture was verified by Alissandrakis, Georgakilas & Dialetis (1992), who found that the waves start out as full circles around oscillating umbral elements with a size of 2-3 and the front of the disturbance propagates in all directions within the umbra; when they reach the penumbra, they propagate in regions with a regular fibril structure.

5 E.B. Christopoulou et al.: Oscillations and running waves observed in sunspots 309 Fig. 5. Time slices image in Hα center, produced using method II; x axis corresponds to the distance from the center of the oscillating element and y axis to time. Tickmarks in the x-axis correspond to 1.3 and tickmarks in the y axis correspond to 140 sec. White lines mark the umbra and penumbra boundaries, as derived from the Fe I 5576 Å images. Fig. 7. Time slices image in Hα+0.35 Å(cf. image 6). In this wavelength rp waves are very weak and we mainly observe oscillations. Notice the reduced amplitude of oscillations halfway the penumbra. Fig. 6. Time slices image in Hα 0.35 Å(cf. image 6). 2. Observations and image processing Observations were obtained at the R.B. Dunn telescope of the Sacramento Peak Observatory with a 512 by 512 pixel CCD camera and the UBF filter. The pixel spatial resolution was A large isolated sunspot was observed at N14.7, E26.0 on August 15, Our observations consist of three sequences of filtergrams. The first one has duration 53 minutes and was Fig. 8. Time slices image from Dopplergrams in Hα±0.35 Å (cf. image 6). Note the prominent 5-minutes oscillations seen well outside the sunspot. obtained in Hα center and ±0.5 Å. The time interval between successive images of the same wavelength was 12 sec. The duration of the second sequence was a few minutes; filtergrams were obtained at 9 wavelengths along the Hα profile (±0, ±0.35, ±0.5, ±0.75, ±1.0 Å). The time interval between successive images of the same wavelength was 36 sec. Finally the third sequence has duration 116 minutes and was obtained in Hα center, Hα±0.35, Hα±0.75, and in the magnetically non sensitive line Fe I ( ) ±0.012 Å. Note that the precision of the UBF

6 310 E.B. Christopoulou et al.: Oscillations and running waves observed in sunspots results are impressive. The running penumbral waves, as well as details inside the umbra are clearly revealed (Fig. 1). Comparing the original image with the processed one in Fig. 1 we verify that the same wavefronts can be observed in both. The only difference is that in the processed image the wavefronts could be better identified. The method that we have used, makes the waves much more noticeable without introducing artifacts. 3. Results 3.1. Observation of running penumbral waves in different layers of the solar atmosphere Fig. 9. Maximum power frequency of the oscillations with distance from the center of the oscillating element, computed from Hα 0.35 Å images. The maximum power frequency corresponds to umbral oscillations in the umbra, is intermixed in the inner part of the penumbra and corresponds to running waves as we move outwards. The outer boundary of the umbra is at about 3 and of penumbra at about 10.6 respectively. filter is of the order of 1 må while the FWHM is about 240 må near Hα and about 120 må near Fe I ( ). The time interval between successive images of the same wavelength was 28 seconds, while the time difference between opposite Hα wings was 4 seconds. We computed Dopplergrams in Hα±0.35, ±0.5, ±0.75, Å as well as in Fe I ±0.012 by red blue wing subtraction. We used the Doppler images as qualitative maps of the line of sight velocity. For the quantitative computation of the velocity more sophisticated methods (that are valid under very demanding conditions) would have been more appropriate. However as pointed out by Alissandrakis et al. (1990), although the Doppler velocities computed with this method are appreciably underestimated, the sign of the velocity is correct. In order to remove the sharp intensity gradient between the umbra and the penumbra and enhance time varying phenomena in intensity images, we applied a subtraction image processing technique. First, we used a cross correlation algorithm in order to position the images with high accuracy. This is very important in order to avoid artifacts. Then, we computed a running local average over a number of images before and after the given image. Finally we subtracted from each image the corresponding local average. The running local average covers roughly one period of the waves in order to produce a background free of waves. Subtracting the local average from the original frame results in enhancing the contrast of intensity modulations produced by the waves (Alissandrakis et al. 1992, Brisken & Zirin 1997, Christopoulou et al. 1999). The filtering technique of the running local average has successfully applied in the whole amount of the images and the We found out that running waves are best observable in the Hα center and in Hα 0.35 Å (Fig. 2). Taking into account the FWHM of the UBF filter this means that they are most prominent near Hα 0.18 Å. In Hα+0.35 Å they are not as prominent as in Hα 0.35 Å. We could detect them mainly near the umbrapenumbra boundary and more difficult near the outer edge of the penumbra, but they were very weak halfway between the inner and the outer edge of the penumbra. This behavior seems not to be related to projection effects or any other kind of asymmetry since the waves are most prominent in Hα 0.35 Å, on both limbward (eastern) and diskward (western) sides of the spot. We could hardly observe the waves in Hα±0.5 Å both intensity and Dopplergrams, and in spite of our image processing technique we could not observe them in Hα±0.75 Å. Our results agree with those presented by Beckers (Giovanelli, 1974). Musman et al. (1976) found in Fe I 5576 isolated wave packets propagating outward across the penumbra. The waves were intermittent and had higher horizontal phase velocity by a factor of 2 or 3, compared to the penumbral ones. We were not able to find obvious running waves in either the photospheric λ 5576 intensity or Doppler images. Our results are in agreement with that of Lites et al. (1982). According to the above authors the failure to find evidence for umbral oscillations in this line does not necessarily indicates that these oscillations are weak in the umbral photosphere, but it is rather due to the fact that the line is weak in sunspots. It seems that the method of observation that we used is not sensitive enough to reveal the waves in this line, if they exist Propagation properties of running waves In order to study running waves we produced movies in Hα, Hα 0.35 Å, as well as Doppler velocity movies in Hα±0.35 Å. Running waves appear to start out from umbral oscillating elements and to propagate outwards forming full concentric circles around the elements (Fig. 3). In Hα center the waves appear to propagate beyond the outer edge of the penumbra, to the superpenumbra. Fig. 4 shows a characteristic example. In the superpenumbra the waves appear more dilute and more diffuse due to their expansion in larger arcs. In order to measure the propagation velocity of the waves, we followed two methods: (i)we drew the wavefronts on transparencies and subsequently we measured the distance between the position of the

7 E.B. Christopoulou et al.: Oscillations and running waves observed in sunspots 311 same front in consecutive images at a number of points along the front. We used this method in order to get rough estimates of the propagation velocity of rp waves. However, it introduces various uncertainties and depends on the observer. Thus we based our analysis on method II, that is considered much more accurate. (ii) We took a cross section of every image of a time series, starting from the center of an oscillating element and directed outwards, at right angles to the propagating wavefronts. We created a new image with distance from the center of the oscillating elements as one axis and time as the other (Figs. 5,6,7 and 8). Figs. 5, 6, 7 and 8 show images computed, using method II, from Hα, Hα 0.35 Å, Hα+0.35 Å and Doppler images in Hα±0.35 Å. The running penumbral waves appear as diagonal streaks, curving forward in time. Apart from the running waves, umbral oscillations are easily discernible as streaks almost parallel to the x axis. The oscillations are not restricted in the umbral area but intrude well into the penumbra. They can be more easily observed in the Hα+0.35 Å image (Fig. 7), where running penumbral waves are very weak. From this image we observe that they are obvious in the umbra and the inner part of the penumbra, almost absent in the central part of the penumbra, and they appear again near the outer edge of the penumbra. Fig. 5 confirms the propagation of rp waves in the superpenumbra. From the distance - time graphs we found that the propagation velocity of rp waves from the inner to the outer part of the penumbra remains constant to about 13 km s 1. In the superpenumbra the propagation velocity seems to remain constant at about the same value as in the penumbra, up to the point where the waves start to dilute. We were not always able to observe rp waves propagating in the superpenumbra; in some cases they seem to end before the outer penumbral edge. However this is probably related to seeing deterioration. Near the penumbra - superpenumbra boundary rp waves seems to be distracted by oscillations. We did not find significant changes for subsequent wave cycles or an obvious long term periodicity. In order to determine the frequency of the waves with distance from the center of the oscillating element, we performed power spectrum analysis. Fig. 9 shows the maximum power frequency with distance from the center of the oscillating element for the Hα 0.35 Å image. We observe that in the umbra dominate frequencies near 6.5 mhz and in the penumbra near 3 mhz. The oscillating frequency is almost constant in the umbra at values reported for chromospheric umbral oscillations. There is an abrupt decrease near the umbra - penumbra boundary; further out the frequency appears to remain almost constant in the central part of the penumbra, where running penumbral waves dominate. Near the penumbra-superpenumbra boundary the results become spurious due to the intermixing of oscillations and rp waves that both are more weak than in the umbra-penumbra boundary. In the inner and middle part of the superpenumbra, where rp waves dominate again, the frequency is almost the same with that in the penumbra Fig. 10 shows the power of the maximum power frequency as it changes with distance from the center of the oscillating elements. The power is high in the umbra, reduces in the region between the umbra and the penumbra, and increases again in Fig. 10. Maximum power and frequency of the oscillations with distance from the center of the oscillating element, computed from Hα 0.35 Å images. the central part of the penumbra. In Fig. 11 we show the power spectra of the points near the umbra penumbra limit. We observe that as we approach the umbra-penumbra boundary (3 )two maxima are present, one near 6 mhz and the other near 3 mhz, which both seem to be real. That is, in this region we observe that two oscillating modes are present corresponding to umbral oscillations and running penumbral waves respectively. Finally in Fig. 12 we show the maximum power frequency with distance from the center of the oscillating element for the Hα+0.35 Å image, where umbral oscillations dominate. In the umbral region frequency is in the 7.5 mhz band. In the penumbra frequencies in the 3.5 mhz band dominate. Finally in the superpenumbra we observe low frequencies near 2 mhz. From the above analysis it is obvious that we observe more than one oscillating modes. There is a fundamental distinction between running penumbral waves and standing oscillations that are observed in umbra, the inner penumbra region as well as in the penumbra - superpenumbra boundary. Oscillations almost disappear in the central region of the penumbra or their magnitude is considerably reduced. Running waves start out near the outer edge of the umbra and propagate up to the outer edge of the penumbra. The propagation velocity is constant from the inner to the outer part of the penumbra. Using method one, we observed an acceleration near the umbra - penumbra boundary and a deceleration near the penumbra - superpenumbra boundary, which however it is possibly spurious. 4. Discussion and conclusions One of the major questions concerning the nature of running penumbral waves is the orientation of the velocity perturbations

8 312 E.B. Christopoulou et al.: Oscillations and running waves observed in sunspots Fig. 11. Power spectra of the points near the umbra penumbra limit computed from Hα 0.35 Å images. At the top of each graph is written the distance from the center of the oscillating element in arcseconds. with respect to the magnetic field vector. Moore(1981) proposed that the mechanism that makes running penumbral waves visible in off-band Hα filtergrams is the Doppler shifting of the Hα, absorption profile in and out of the filter bandpass by periodic vertical mass motions. Giovanelli (1974) reported no success in his search for oscillatory motions in penumbrae observed in Hα near the limb. He further failed to find waves using a filter with 0.7 Å pass band centered on Hα; with this pass band, pure Doppler shifts without an associated change in depth of the line should produce no change in the penumbral intensity. Based on the above results he proposed that the observed disturbance is transverse and not longitudinal. On the other hand Maltby (1975) interpreting Hα observations of two sunspots located at about 35 from disk center, and based on that they had undetectable wave activity on their sides away from disk center proposed that velocity disturbances are aligned with the magnetic field. Furthermore, Lites (1992) reported that the attenuation ring that he observed in the Fe I 5434 Å velocity images (cf. introduction) has its closest approach to the umbra on the side of the sunspot away from disk center. Based on the above result he proposed that the fluid motion associated with the penumbral oscillations is primarily aligned with the magnetic field in the high photosphere of the inner penumbra. Our observation that the penumbral waves are much more prominent in Hα 0.35 Å, indicates that the penumbral perturbation is not symmetric about zero, but primarily upward. From the Dopplergrams we found out that the oscillation of the umbral elements is symmetric about zero. However we could not verify if the perturbation related to rp waves is primarily upward due to the superposition of the Evershed flow. In case the above indication is approved to be correct then, if the perturbations were aligned parallel to the magnetic field, the waves should be more obvious in one wing at the diskward side of the spot, and in the other at the limbward side of the spot. However, the waves are much more prominent in Hα 0.35 Å, in all directions. Thus we conclude that the penumbral perturbations related to rp waves are primarily aligned vertically to the magnetic field. Recently a number of authors (cf. introduction), found either a decrease of the propagation velocity of rp waves as they move outwards or a decrease of the frequency of oscillation as we approach the outer edge of the penumbra. Possible explanation proposed were the following: (i) The decrease is due to a change of the physical properties with distance from the center umbra (strength of the magnetic field, geometry of the magnetic field lines, density, etc.). (ii) Zhugzhda & Dzalilov (1984) proposed that the decreased visibility and propagation velocity of the penumbral waves in the outer penumbra might be due to the reverse (inflowing) chromospheric Evershed flow. Running waves propagate against the flow and the propagation velocity decreases as we move outwards. Lites (1988) found that the disturbances in the chromospheric line Ca II 8498 Å and in the high photospheric Fe I

9 E.B. Christopoulou et al.: Oscillations and running waves observed in sunspots 313 Fig. 12. Maximum power frequency of the oscillations with distance from the center of the oscillating element, computed from Hα+0.35 Å images. In Hα+0.35 Å, running penumbral waves are very weak and the oscillation frequency corresponds mainly to umbral oscillations. The outer boundary of the umbra is at about 3 and of penumbra at about 10.6 respectively Å are cospatial. He argued that since in the Fe I 5434 Å the Evershed motion has its familiar property of outflow, would tend to increase the propagation velocity of the waves. Further the reported large scale average velocities of the Evershed flow, do not approach sonic speeds and do not justify the observed deceleration. Fe I 5434 Å line is formed near the temperature minimum and there are no reports concerning the Evershed outflow velocity. Dere et al. (1990) found Evershed radial outflows of up to 1.1 km s 1 in the photosphere and inflows of up to 2.6 km s 1 in the Hα±0.35 Å chromosphere (mach numbers of about 0.15 and 0.25 respectively). In the chromosphere the velocity peaks at a greater distance than in the photosphere. According to Dialetis et al. (1985) the maximum of the velocity in the photosphere is above the penumbra, while in the chromosphere it is well outside the penumbra. A possible scenario then, is that the waves are slightly accelerated from the photospheric outflow up to a point in the penumbra and then they are decelerated by the chromospheric inflow near the outer edge of the penumbra. However we observed that waves propagate with a constant velocity, thus the influence of Evershed flow should not be considered significant. (iii) Finally the third explanation proposed is that we observe more than one modes of oscillations and that the decrease of the oscillating frequency is due to a switch in the dominant mode of oscillation. Our results support this explanation, since we observed more than one oscillating mode. Running penumbral waves, that have a constant velocity and oscillation frequency in the penumbra and the superpenumbra, umbral oscillations and oscillations in the penumbra-superpenumbra boundary. A number of authors have proposed that the reduced amplitude of the oscillations in the central part of the penumbra and the decreased frequency of the oscillations near the penumbra-superpenumbra boundary are due to the p-mode influence becoming more important (Lites, 1992; Abdelatif et al. 1986). The magnitude of the oscillations seems to change with distance from the center of the umbra. Thus, when they are near regions of abrupt change of intensity (umbra - penumbra boundary) they appear as extended wavefronts and are confused with running waves. Further if the time slices analysis is performed to lower resolution observations, then it is easy to confuse the two modes. In such a case waves appear to have large velocities near the umbra penumbra boundary and to decelerate as they move outwards. Another question that has not yet found a definite answer is the relation of umbral oscillations with running waves. From the time slice images it is obvious that there is not a clear relationship between umbral oscillations and running penumbral waves; rp waves are not the continuation of umbral oscillations into the penumbra. The period of rp waves is about twice as that of umbral oscillations. A possible explanation is that the different period is due to a change of the physical properties with distance from the center of the umbra. Further, small or vanishing phase velocities near the umbra-penumbra border may be caused if the waves are propagating primarily vertically there (propagation towards the observer ). The fact that umbral oscillations intrude well into the penumbra, and the fact that the frequency of rp waves remains constant in the penumbra do not favor these explanations. Moreover, near the umbra penumbra boundary, the two oscillating modes are clearly distinguished and there is not a smooth transition from the one to the other (Figs. 6, 12). If we accept that the rp waves are generated by umbral oscillations or by the same cause with umbral oscillations, then a mechanism explaining the different period is necessary. Another possible explanation is that the waves are related with the 5 minutes oscillations present in the umbral photosphere. This is a subject that we will attempt to investigate more extensively in a subsequent work. In order to further investigate the nature of oscillations and waves related to sunspots, more sophisticated observations of higher sensitivity are needed as well as parallel analysis of other penumbral phenomena, like the Evershed flow. Simultaneous observations in several lines at different depths of the chromosphere and photosphere will help to determine how deeply sited the oscillations and the waves are. Acknowledgements. We would like to thank Dr. R. N. Smartt, the T.A.C. of NSO/SP and the staff of the Sacramento Peak Observatory for their warm hospitality and their help with the observations. Further we would like to thank the referee (Dr. Wittmann) for his helpful comments. References Abdelatif, T.E., Lites B.W., Thomas J.H., 1986, ApJ 311, 1015 Alissandrakis C.E., Tsiropoula G., Mein P., 1990, A&A 230, 200 Alissandrakis C.E., Georgakilas A.A., Dialetis D., 1992, Solar Phys. 138, 93 Bhatnagar A., Tanaka K., 1972, Solar Phys. 24, 87

10 314 E.B. Christopoulou et al.: Oscillations and running waves observed in sunspots Bhatnagar A., Livingston W.C., Harvey J.W., 1972, Solar Phys. 27, 80 Beckers J.M., Schultz R.B., 1972, Solar Phys. 27, 61 Brisken W.F., Zirin H., 1997, ApJ 478, 814 Christopoulou E.B., Georgakilas A.A., Koutchmy S., 1999, ASP Conference Series, vol. 184, p. 103 Chitre S.M., 1992, In: Thomas J.H., Weiss N.O. (eds.) Sunspots, Theory, Observations. Kluwer, Dordrecht, p. 333 Dialetis D., Mein P., Alissandrakis C.E., 1985, A&A 147, 93 Dere K.P., Schmieder B., Alissandrakis C.E., 1990, A&A 233, 207 Galloway D.G., 1978, MNRAS 184, 49P Giovanelli R.G., 1972, Solar Phys. 27, 71 Giovanelli R.G., 1974, In: Grant Athay R. (ed.) IAU Symposium 56, Chromospheric Fine structure. p. 137 Kentischer T.J., Mattig W., 1995, A&A 300, 539 Lites B.W., White O.R., Packman D., 1982, ApJ 253, 386 Lites B.W., Thomas J.H., 1985, ApJ 294, 682 Lites B.W., 1988, ApJ 334, 1054 Lites B.W., 1992, In: Thomas J.H., Weiss N.O. (eds.) Sunspots, Theory, Observations. Kluwer, Dordrecht, p. 261 Maltby P., 1975, Nat 257, 468 Musman S., Nye A.H., Thomas J.H., 1976, ApJ 206, L175 Moore R.L., Tang F., 1975, Solar Phys. 41, 81 Moore R.L., 1981, Space Sci. Rev. 28, 387 Shröter E.H., Soltau D., 1976, A&A 49, 463 Soltau D., Shröter E.H., Wöhl H., 1976, A&A 50, 367 Sigwarth M., Mattig W., 1997, A&A 324, 743 Thomas J.H., Weiss N.O., 1992, In: Thomas J.H., Weiss N.O. (eds.) Sunspots, Theory, Observations. Kluwer, Dordrecht, p. 3 Zhugzhda Y.D., Dzhalilov N.S., 1984, A&A 133, 333 Zirin H., Stein A., 1972, ApJ 178, L85