No abstract

No abstract

We present a comprehensive set of spectral data from two quiescent solar prominences observed in parallel from space and ground: with the VTT, simultaneous two-dimensional imaging of Hβ 4862Å and Ca ii 8542Å yields a constant ratio, indicating small spatial pressure variations over the prominence. With the Gregory, simultaneous spectra of Ca ii 8542Å and He i 10830Å were taken, their widths yielding 8000 < T kin < 9000 K and v nth < 8km/s. The intensity ratio of the helium triplet components gives an optical thickness of τ < 1.0 for the fainter and τ ≤ 2.0 for the brighter prominence. The τ 0 values allow to deduce the source function for the central line intensities and thus the relative population of the helium 3 S and 3 P levels with a mean excitation temperature T mean ex = 3750 K. With SUMER, we sequentially observed 6 spectral windows containing higher Lyman lines, 'cool' emission lines from neutrals and singly charged atoms, as well as 'hot' emission lines from ions like O iv, O v, N v, S v and S vi. The EUV lines show pronounced maxima in the main prominence body as well as 'sidelocations' where the 'hot' lines are enhanced with respect to the 'cool' lines. The line radiance of 'hot' lines blue-wards of the Lyman series limit (λ < 912Å) appear reduced in the main prominence body.This absorption is also visible in TRACE images of Fe ix/x 171Å as fine dark structure which covers only parts of the main ('cool') prominence body.The Lyman lines show a smooth decrease of line widths and radiance with increasing upper level k = 5 through 19. For 5 ≤ k ≤ 8 the level population follows a Boltzmann distribution with T ex > 6 · 10 4 K; higher levels k> 8 appear more and more overpopulated. The larger widths of the Lyman lines require high non-thermal broadening close to that of 'hot' EUV lines. In contrast, the He ii emission is more related to the 'cool' lines.

The time dependence of Doppler shift and line-center intensity is simultaneously observed for the Hc~ emission of three solar prominences, each one during about two hours. Doppler oscillations with periods near one hour and amplitudes between 1 and 2 km s ~ are conspicuously visible in the recordings of all three prominences. Fourier analysis yields periods of 50, 60, and 64 min, as well as slight indications of short periods near 3 and 5 rain. No oscillations are found in the line-center brightness.The time variation of the displacement of the Ha emission was measured with the evacuated Gregory telescope at Locarno of the Grttingen Observatory (Wiehr et al., 1980) on October 1, 3, and 4, 1983 in three prominences at the NW, SW, and NW limb, respectively. An entrance hole of 8 arc sec diameter was centered on a selected area of each prominence and guided for about two hours. Using the Zeeman polarimeter (Wiehr, 1969) three spectrograph slits covered the Ha emission profile. The central-line intensity was measured through the central slit at 20 + 50 rnA, the Doppler plate was driven by the signals from the neighbouring slits at + (50 to 250) mA. Calibration of the line shift was performed turning the grating by a known amount. Both line-of-sight velocity and intensity signals were stored on magnetic tape and then subjected to a fast Fourier transform analysis (e.g. Brault and White, 1971). Edge discontinuities were not removed. For the slowly varying intensity recordings, a quadratic trend was subtracted before.The velocity recordings in Figure 1 show for all three prominences Doppler oscillations of the Ha amplitudes between 1 and 2 km s -1. The velocity power spectra Figure 2 confirm these results yielding defined periods of 60, 64, and 50 min :for the October 1, 3, and 4 prominences, respectively. This result roughly confirms the finding by Bashkirtsev et al. (1983) of 77 and 82 min periods. Figure 2 additionally indicates short-period oscillations of lower power. If the time coherence of these oscillations is short as compared to our two-hour samples, i.e. if they last only a few periods, their maximum power should increase when studying subsamples of shorter time intervals. Such a behaviour is particularly found for the 3-min period from October 1 which exists only in the last 40-min time-interval of that sample, yielding a power maximum of 4.5 • 104 (m s-1)2 corresponding to an amplitude of