One of the basic astrophysical problems is the search for and investigation of brightness fluctuations of celestial objects on time scales comparable with the mean time between registered photons. To facilitate this a special mathematical formalism was developed (Shwartsman 1977, Plokhotnichenko 1983).

At the Special Astrophysical Observatory, photometric equipment has been developed which allows these ideas to be explored observationally and which also allows to study light curves of astrophysical objects using classical methods.

This equipment and these methods are used in the search for and study of variable objects with time resolution 10−7 s in the MANIA (Multichannel Analysis of Nanosecond Intensity Alterations) experiment. The hardware consists of a photon-counting photometer for synchronous detection in different colour bands, a specialized “time-code” converter (Quantochron 3-16), measuring photon arrival times with an accuracy of 20 ns and a PC AT 386 and DAT-cassette recorder. Our equipment allows the uninterrupted accumulation of 108 photocounts with a rate up to 375 kHz without distortions in 28 or 216 parallel channels (Zhuravkov et al. 1994). The software permits the investigation of variability on time scales 10−7 − 102 s.

A new prefiltering technique (PFT) for precise stellar photometry is described. It has been proved that the prefiltered and added high-time resolution data provide an increase in S/N ratio by factors 1.3 — 1.7 in comparison with simple integration. The PFT allows to detect event amplitudes which are comparable to that caused by counting statistics. Adopting the PFT, our observations reveal small-scale flare events in the subsecond range in EV Lac.

Early studies of stellar flares were made entirely in the optical regime. It was recognised that flares arose from the generation of hot plasma within the stellar chromosphere at whose temperature (indicated, for instance, by the presence of a strong, blue optical continuum) a substantial emission in the ultraviolet would be expected. It was not until the advent of space-borne instruments of adequate sensitivity, however, that direct confirmation of this prediction was forthcoming. In this review I examine some results of more than a decade of observation of stellar flares in the UV.

The major source of ultraviolet data on stellar flares has been the International Ultraviolet Explorer (IUE) satellite (Boggess et al. 1979). In order to understand the limitations of our current understanding in this area it is important to appreciate some of the characteristics of its instrumentation. IUE’s telescope is of 40 cm aperture and it is equipped with a spectrograph which can operate at two resolutions, i.e. Δλ/λ ∼350 (LORES) and ∼17000 (HIRES). Its detectors are optimised for operation in two wavebands, i.e. ∼1150-1950Å (SW) and ∼1950-3200 Å (LW). IUE’s small apert ure results in a limited sensitivity, a consequence of which is a modest time resolution when studying stellar flares (a long exposure time is needed to gain adequate signal-to-noise). IUE’s elliptical 24-hour quasi-geosynchronous orbit and its resulting interactive mode of operation make continuous monitoring feasible, a feature suiting flare star work.

Data of long-duration emission arising after the impulsive rise and decay in a flaring event on the red dwarf star AU Mic are discussed. Intensive EUV emission in the band 65-190 Å was registered by the Extreme Ultraviolet Explorer (EUVE) after both impulses during half a day. A similar behavior of the flux in the Fe XVIII 93.9 Å line is detected after the first powerful impulse. The decay of the intensity in the 65-190 Å band and in the Fe XVIII line during this prolonged event is 10 times slower than the time of radiative cooling of coronal loops with a typical flare plasma density. Some difficulties with two explanations of this event proposed earlier are discussed. Explanation (i) - the radiation of dense loops at main phase of the flare, and (ii) - the emission of the low-dense plasma of coronal transients (CME). The temporal behavior of the emission measure is determined for both the 65-190 Å band and the Fe XVIII line fluxes. The total energy emitted in the 1-2000 Å region for the long-duration event lasting almost 12 hours is 3 · 1035 ergs. The following physical model is proposed to explain the prolonged event (Fig. 1): the source of emission is the system of high coronal loops, the size of which is more than the active region scale, but is less than the stellar radius. The temperature of the plasma in the loops decreases from 107 K slowly, during a few hours. The densities in these loops are in the range 1013 cm−3 to 5 · 109 cm−3. Such systems, when the plasma therein becomes cool, are observed in the Hα line during large solar flares (for instance, June 15, 1991) after CME. Some additional post-flare energy input into this high coronal loop systems can be caused by the reconnection in vertical current sheet, and this post-eruptive energy release provides prolonged and intensive EUV emission.

Apparently, we are faced here with a new kind of surface activity on late-type stars which is intermediate between impulsive flares on red dwarfs and long-duration, powerful events on subgiants, which are components of RS CVn binaries.

The full version of this contribution will be published in Astronomicheskij Zhurnal (Astronomy Reports), 1995, Vol. 72.

Flare-like brightenings, similar to those observed on the Sun but on a larger energy scale, are observed in a variety of late-type stars, including classical M dwarf flare stars, RS CVn-type binaries and pre-main sequence (PMS) objects. These events, which are observable over the whole electromagnetic spectrum, are interpreted as due to rapid release of magnetic energy at the star surface. Comprehensive reviews of stellar flares have been presented by Haisch &: Rodonò (1989), Pettersen (1991), Haisch et al. (1991), Pallavicini (1992a), as well as in contributions at this conference. Here I will discuss X-ray emission, with emphasis on flares on M dwarfs and RS CVn binaries.

X-ray observations are important because they provide information on the high-temperature coronal portion of the flare and, at least in principle, could give clues as to the primary energy release and particle acceleration. In the stellar case there are severe limitations of the diagnostic capabilities of X-ray observations, in comparison to the solar case. In fact, stellar flares have up to now been observed only at soft X-ray energies (≤ 10 keV) where thermal processes greatly dominate over non-thermal ones (optical continuum observations have to be used in this case as a proxy for hard X-rays). Moreover, stellar X-ray observations are spatially unresolved and so far of poor spectral resolution. Thus, the observational data that are readily accessible are only light curves, average temperatures T and volume integrated emission measures EM = ∫ n2dV, but there is no direct information on parameters such as density, volume and detailed geometry.

During 1990-92, the WATCH all-sky X-ray monitor on GRANAT has discovered 6 short-duration X-ray transients. We discuss their possible relationship to peculiar stars. Only one source, GRS 1100-77 seems to be related to a T Tauri star.

ROSAT/PSPC observations of a long-duration, giant X-ray flare on the eclipsing binary Algol are presented. With a total thermally radiated energy of 7 Ŀ 1036 erg, this flare is one of the largest events ever observed. The flare decay is modeled by means of the quasi-static cooling method, which yields a height of the flare loop arcade of ≃ 1RK, and a plasma density of 1011 cm−3.

The Japanese satellite ASCA carries X-ray detectors which combine excellent sensitivity and high spectral resolution. We observed the young, rapidly-rotating K dwarf AB Doradus with ASCA in November 1993. The star’s X-ray flux was essentially steady for the first half of the observation, and then a series of flares occurred during the second half. The flares showed rise times of 30 minutes and decay times of several hours. The quiescent X-ray luminosity was 2 × 1030 ergs s−1 (0.5-10 keV). At the flare peaks, LX(> 0.5 keV)/Lbol = .002.

AB Dor has a relatively high count rate in the ASCA data and we can obtain high-quality spectra for different periods of the light curve. In Fig. 1 we present spectra corresponding to the period of quiescence, the rising phase and peak of each of the three flares combined, and the decay periods of each of the flares combined. The rise/peak curve (upper) has been multiplied by 2 for purposes of display. We have subtracted the quiescent spectrum from the flare spectra (i.e., used it as a background spectrum) for this display and subsequent analysis.

The detailed differences between these spectra display the importance of the high spectral resolution provided by ASCA. One important difference, not readily evident in this figure, is the Fe K line complex at 6.4-6.7 keV. This line, a diagnostic of hot plasma, is strong in the rise/peak spectrum, moderate in the decay spectrum and weak in the quiescent spectrum. A time profile of the counts in this energy range shows little significant emission during the quiescent periods but strong emission at the flare peaks.

The flare/spotted spectroscopic binary star CC Eri (HD 16157) was observed in the periods 1990 July 9-11 and 1992 January 26-27 with the ROSAT PSPC detector (Pan and Jordan 1995). These high quality data give the first information on the temporal and spectral variability of CC Eri in the X-ray energy band.

During the ROSAT observations the X-ray intensity of the source is variable on timescales from a few minutes to several hours. The X-ray luminosity is in the range 2.5 – 6.8 × 1029 erg s−1, which is similar to values found from previous measurements with the Einstein IPC and EXOSAT LE. On 1990 July 10 an X-ray flare-like event was detected with an exponential rise time of about one hour and a decay time of about two hours. The X-ray spectrum of CC Eri can be described by current thermal plasma codes with two temperature components or with a continuous temperature distribution. The spectral results show that plasma at Te ∼ 107 K exists in the corona of CC Eri. We find that the X-ray spectrum is also variable and the variations of the emission measure, and to a lesser extent of the temperature, are correlated with the source intensity.

During Ginga observations of galaxy 3C390.3 on 1991 February 14, a large impulsive flare was detected between 22:30-23:00 UT (Inda et al. 1994). Inda et al. (1994) suggested that the flare was associated with EQ1839.6+8002, an M4Ve star at a distance of 15.2 pc (Fleming et al. 1988). The star has been observed previously with the Einstein IPC and EXOSAT LE (Gioia et al. 1990; Grommi et al. 1991), and more recently with the ROSAT PSPC.

In this paper we present the results of an analysis of the observations of EQ1839.6+8002 with Ginga and of the data obtained with the Einstein IPC (between 1980 January and May), the EXOSAT LE (in 1984-1986), and the ROSAT PSPC instruments (in 1991 March, 1992 April and 1993 April).

Investigations of impulsive flares on both the Sun and red dwarf stais during more than 30 years allow us to arrive at quite definite conclusions. Here we will consider impulsive events; on the Sun the impulsive phase of a flare is observed as a hard X-ray burst with the emission of photons with energies E > 30 keV up to the γ-ray range. At the same time microwave radio bursts, and sometimes UV and optical continuum bursts are registered. Typical durations of these processes are ∼l-3 min. In this time interval other kinds of flare emission like soft X-ray (2-10 keV) emission, meter radio bursts and Balmer line emission begin to rise, but their maxima occur later on, in the gradual (thermal) phase of the flare.

Impulsive stellar flares are often observed as a significant increase in optical continuum, especially in the U-band, of similar duration (1-3 min), and this time interval is, like in the solar case, the rise phase of the soft X-ray emission.

Modern observations demonstrate that both the impulsive phase of a flare or an impulsive flare develops in low-lying loops. Earlier only indirect evidence existed in optical and radio data. Recently, however, the heights of the hard X-ray sources in impulsive solar events were determined directly from YOHKOH’s HXT (Kosugi 1994, Masuda 1994) (Fig. 1a). Statistically, the height of the hard X-ray source in the 14-23 keV range is 9700 ± 2000 km above the photosphere, and this height reduces to 6500 km in the 53-93 keV range. Besides two hard X-ray sources in the loop footpoints, a third hard X-ray source exists at the top of the loop at least in some cases. The authors of this experiment suppose that the appearance of this loop-top source is due to reconnection in the impulsive phase. Note that the reconnection begins close to the apex of the loop, when this loop is filled by hot plasma that evaporated from both footpoints.

Supernovae and their expanding shock fronts are evidently the main agents of cosmic-ray acceleration. The thermal gas in the interstellar medium has been regarded as the reservoir of seed particles destined to become cosmic-ray nuclei. This assumption is, however, at variance with the source composition of galactic cosmic iays. In an alternative hypothesis, the seed particles are injected into the interstellar material as suprathermal seed ions, and it has been surmised that flare stars provide the initial boost. We find that the dMe and dKe stars are probably the principal sources of cosmic-ray seed particles. Most stars in the Galaxy are red dwarfs and many of these flares much more powerfully and frequently than solar flares. Augmenting the optical data, recent X-ray and far-ultraviolet observations now permit a better estimate of the energy budget. Altogether, dMe and dKe stars seem to be the most promising class of cosmic-ray injectors.

Two sites of flare activity have been identified in Algol-type binaiies. One is associated with the shock region from the accretion on B-A type primary, another, predominant one is apparently connected with coronal mass ejections (CME) from the late-type, Roche lobe filling secondary. The relative contribution from both sources of activity is still to be found. According to Peters & Polidan (1984), a moderately hot (105 K) and low density ne ≃ 109 cm−3 high turbulent accretion region (HTAR) surrounding the hot component has been discovered. Typical dimensions of HTAR agree with the scale height of the X-ray source HX ≃ 1.2 · 1011 · T7 cm (Harnden et al 1977), where T7 is the temperature in units of 107 K. But the decay time of the HTAR, comparable to the orbital period, more favourably agrees with the idea of CME as deduced from Ginga observations of X-ray flares on Algol (Stern et al. 1992). One solution of the controversy as to the contribution from HTAR would be to concentrate on X-ray observations of systems like u Her and U CrB, where the secondary is possibly earlier than FO (thus, is not expected to possess an extensive convection zone) and on the other hand on systems like U Cep, where during periods of high activity Ṁ ≃ 10−6M⊙ per year. Thus, the accretion luminosity must be at least by two orders of magnitude higher than in Algol. At present the analysis of X-ray flares is based upon the solar type CME assuming different scenarios of radiative and/or conductive cooling (see Stern et al. 1992). However, the current models of CME do not take into account the binary nature of the flaring object.

Low-mass main sequence stars show a magnetic activity similar to the Sun and as a consequence they lose mass in the form of a variable stellar wind. In the latest spectral types (red dwarfs) the activity and the mass loss rate appear to increase by a large factor of ∼ 103 with respect to the solar case, reaching Ṁ ∼ 2 · 10−11 M⊙/yr (Badalyan & Livshits 1992, Katsova 1993). The same happens for coronal mass ejections (CMEs), which are the most relevant transient events of mass loss in these objects. In the Sun, these appear as bubbles of coronal material, with dimensions of some fraction of the solar surface, mass M ≃ 2 · 1014 - 2 · 1016 g and ejection velocity v ≃ 3 · 107 - 2 · 108 cm/s, with an instantaneous mass loss rate Ṁ ∽ 10−13 - 10−11 M⊙/yr (Wagner 1984). In red dwarfs, as recently observed, the ejection velocities are higher, up to v ≃ 3 · 108 cm/s and the mass loss rate can reach the value Ṁ ≃ 10−8 M⊙/yr (Mullan et al. 1989, Houdebine et al. 1990). In both cases, the observations suggest that a bubble expands, once ejected, with a velocity of several hundreds of km/s.

Recent observations of the EGRET-telescope aboard the Compton Gamma Ray Observatory (CGRO), results from Čerenkov shower telescopes, and fast variability studies have demonstrated the dominant role of nonther mal processes for some accreting relativistic objects. This is perhaps an unex pected result, because the basic assumption of the standard accretion models (Shakura & Sunyaev 1973, Sunyaev & Titarchuk 1980, Galeev et al. 1979) is the thermal nature of all processes that determine the energy output. At the same time, the recent observations argue that the source of the emission is not the thermal plasma as a whole, but a very small fraction of charged particles accelerated to very high energy (ε/mc2 > (30-1000) with a “universal” energy spectrum nε ∣ ε−α. These observational results suggest a revison of the global approach to the accretion process to find any new (or forgotten “old”) energy conversion mechanisms with predominately nonthermal output (Lynden-Bell 1969, Ikhsanov & Pustil’nik 1994).

The nature of the energy source and mechanisms of energy conversion to either plasma heating (“thermal” mode) or to acceleration of charged particles (“nonthermal” mode) is the main question for accretion theories.

This review concerns rapid variability, on time-scales of a few hours or less, occurring on T Tauri stars (TTS). There are several recent reviews on observed properties of TTS (e.g. Appenzeller & Mundt 1989, Bertout 1989, Gahm 1990a, Kuhi & Cram 1989), some with more emphasis on “flare-like” activity (Feigelson et al. 1991, Gahm 1990b, Montmerle 1991, Montmerle et al. 1993), and the ambition below has been to summarize more recent work. We will also give a compilation of published works on observations of optical variability on these time-scales and a related statistical overview. Some early, very interesting and extensive studies of this kind were made here at the Sonneberg Observatory, where for instance Götz & Wenzel (1967) concluded that the light-curve of RW Aurigae contains several components: quasi-periodic fluctuations over days, ‘waves’ over several hours, rapid outbursts with symmetric light-curves, rapid outbursts with asymmetric (flare-like) light-curves (see also Fürtig & Wenzel 1964) and small fluctuations with very small amplitudes, which they postulated could be due to changes in the emission line fluxes. Modern photometric results of higher sensitivity and time-resolution confirm the existence of these different types of rapid variations in classical TTS (CTTS), having strong emission (lines and continuous, so called veiling) superimposed on the photospheric absorption line spectrum. As described in the subsequent chapters there are new concepts for the interpretation of the short-lived fluctuations. Concerning the long-term quasi-periodic variations we can now usually relate them to the stellar rotational period (bright or dark spots), let be that RW Aur still is an uncertain case. There are other slow changes most likely related to variable circumstellar extinction in circumstellar dust in the line-of-sight to the star. The long-term optical changes were most recently discussed by Herbst et al. (1994).

The interest in the giant flares found on young stars results from the fact that their energetics is vastly different from solar flares. The recent ROSAT observations of young clusters yield a good basis for a systematic study of these flares.

We have investigated 18 very deep pointed observations of star forming regions in IC348, Chamaeleon, ρ Ophiuchi, and Orion. To study the age dependence of the flare properties we also investigated the observations of the young clusters IC2391, Pleiades, and Hyades. In this way we have evaluated the X-ray emission of several hundred young stars with ages ranging from about 106 years to 7 ń 108 years. We have found 36 large flares on identified cluster members and have determined the flare properties with a simple flare model.

The flares show the same relation between the total X-ray flare energy and the quiescent X-ray luminosity as the dM star flares. The total X-ray flare energies decrease smoothly from the very high values found for very young stars to solar values with increasing cluster age. Our results indicate that the large flares on young stars are similar to scaled-up solar-like activity rather than to accretion processes.

We study the X-ray emission of several hundred (young, low-mass, late-type, pre-main sequence) T Tauri stars (TTS) in the Taurus T association, a nearby well-studied region of ongoing star formation. We report on X-ray emission variability of TTS as observed with the flux-limited ROSAT All-Sky Survey (RASS). Since RASS observations are spatially unbiased, we can investigate the X-ray flare rate of TTS on a large sample. We find that large flares are very rare (once per year), while medium-size flares can occur once in ∼ 40 days.

Light variations of different kinds appear to be one of the characteristic features of T Tau stars. They are mainly caused by changes in the continuum radiation. The availability of large observational databases makes it possible to investigate different patterns of variability such as stochastic, quasiperiodic and quasicyclic ones with typical time scales from days to years, as well as sporadic flare and fuorlike outbursts. Recent attempts in classification of the photometric variability of PMS-stars were made by Herbst (1994). In this paper, we study the observational results on flares in T Tau and related stars belonging to the Taurus-Auriga complex, which resemble that of ordinary flare stars.

A census of T Tau stars with homogeneous observations of short and middleterm flares is presented in Table 1. Our selection criteria were: pre- and postilare levels of brightness ought to be quasiquiescent; the amplitude has to be more than 3 times the value of standard error; the duration of brightening does not exceed a few hundred minutes; the lightcurve shape should be similar to that of a typical flare. For convenience, all observational data in Table 1 are divided into 3 subgroups.

We have made a detailed investigation of the short-term variability of the classical T Tauri star BP Tauri in UBVRI. Data were collected from the Wendelstein Observatory in 1991, 1992 and 1993 with time resolutions down to 1 sec. The 0.8m telescope was equipped with a fiber-fed fifteen channel high-speed photometer (Barwig et al. 1987). Observations (in UBV) were also collected in China at the Yunnan and Shanghai Observatories to get a long base line in time. To search for differences in the properties between the brightness variations of classical T Tauri stars (CTTS) and T Tauri stars with weak emission lines (WTTS) we performed simultaneous photometry (in the UBV and Strömgren systems) and spectroscopy of 6 young stars during two observing periods at ESO La Silla. The study concerned mainly short-term variability on time-scales of minutes to a few hours. The sample contained two CTTS, SY Ori and VW Cha; three WTTS, San 1, SZ Cha and ADA 481 and one post-T Tauri candidate, HD 70309B.