With the launch of SOHO new opportunities have become available for studying chromospheric dynamics and, especially, chromospheric oscillations at UV wavelengths. The study of chromospheric oscillations is to a large extent based on optical observations (see e.g. review by Rutten, 1995), while at ultraviolet wavelengths the number of studies is relatively limited. At the end of the seventies several studies of chromospheric oscillations were made based on OSO 8 data (Athay and White, 1979ab, White and Athay, 1979ab, Chipman, 1978, Bruner, 1978, 1981). The OSO 8 data made it possible to study oscillations both in the upper chromosphere (Fe II, Si II, C II) and the transition region (C IV). Although the results of these studies are sometimes conflicting, it could be established that the presence of sound waves is relatively common in the chromosphere, while in the transition region sound waves are not always found. No evidence was found for the existence of a chromospheric cavity as was predicted by Ando and Osaki (1975). An important result of the OSO 8 studies is that the energy flux carried by acoustic waves in the chromosphere is too low to account for coronal heating. For further discussion on chromospheric and coronal heating via acoustic waves, see Narain and Ulmschneider (1996).
In the low chromosphere oscillations with periods of 180-240 s are found (Deubner, 1981) although it is not clear whether these are propagating or evanescent. For example, Lites and Chipman (1979) conclude that in the lower chromosphere 5 minute oscillations are evanescent while waves with frequencies above 5 mHz (200 s) are propagating.
For the upper chromosphere Chipman (1978) found that not all observations show evidence of oscillations but that, when oscillations are present, these have periods near 300 s. Chipman concludes that the associated waves are evanescent. Some of Chipman's results are at variance with those of Athay and White (1979ab) and White and Athay (1979a) who studied a pair of Si II lines. These authors conclude that periods near 300 s are almost always present. The power spectra (line intensities and Doppler shifts) have broad maxima, in the 2.5-9 mHz range, which are superimposed on a flat noise continuum which extends up to 30 mHz and contains most of the power. The waves near 300 s seem to propagate vertically. In the 2.5-9 mHz range delays in the oscillations of the lines, which are formed at different heights, are consistent with upward propagating waves although the phase differences between intensity variations and velocities do not conform to those expected for sound waves.
For the transition region Athay and White (1979b) find that low-amplitude aperiodic fluctuations characterize the data (C IV line) and that only 20% of the datasets show periodic observations in the 3-5 mHz range. These authors argue that the OSO 8 C IV results are indicative of the presence of sound waves in the transition region whose periodicity is sometimes destroyed due to propagation through the chromosphere. Similar conclusions were reached by Bruner (1978) who also found (Bruner, 1981) that the acoustic fluxes in the up- and downward direction almost balance.
OSO 8 had a spatial resolution of . An improvement of the spatial resolution was achieved with the launch of the Ultraviolet Spectrometer and Polarimeter (UVSP) on board the Solar Maximum Mission (SMM). Surprisingly the number of oscillation studies with UVSP is relatively limited. A number of studies focussed on transition region oscillations above sunspots (Gurman et al., 1982, Henze et al., 1984, Thomas et al., 1987), using mainly the C IV line. Bruner and Poletto (1984) conclude from N V (1238Å) observations that there is a net upwardly directed acoustic flux (contrary to Bruner, 1978).
From a rather arbitrary sample of (continuum) light curves of active regions Drake et al. (1989) conclude that periods in the range 4-5 min. (3.3-4.2 mHz) are ubiquitously present while occasionally periods down to 3 min. (5.6 mHz) are seen.
In this paper we discuss the UVSP observations of a number of active regions in order to extend the study by Drake et al. (1989). Although UVSP does not have the diagnostic capabilities of e.g. SUMER and CDS on board SOHO, it is interesting to see which periodicities are present in the UV.
Any analysis of oscillatory phenomenon is necessarily based on Fourier techniques. For observations based on individual photon counts the problem of noise is especially noteworthy. It is important to establish the statistical significance for any feature in a power spectrum resulting from a Fourier transform. Fortunately, this problem has received considerable attention in the X-ray community working on QPO phenomenon. Based on that work we discuss in this paper several effects of noise in power spectra and how the noise levels can be determined. The described techniques are equally applicable to SUMER and CDS data.
The outline of the paper is as follows. In Sect. 2we describe the observations and the spectral properties of the UVSP bandpass. In Sect. 3the method of analysis is described which is applied to the UVSP data in Sect. 4. Our conclusions are presented in Sect. 5.
© European Southern Observatory (ESO) 1997
Online publication: April 8, 1998