next previous
Up: A three-year Strömgren photometric stars


Subsections

4 Results and discussion

Will be presented and successively discussed in that section the three different types of variations detected during our survey.

4.1 Comparison with the Geneva photometric survey

In parallel to our main Strömgren program, the Geneva survey went on with the 70 cm Swiss telescope and the P7 UBVB1 B2 V1G photometer at La Silla Observatory until its definitive end in 1997 (cf. Table 2). Figures 4 and 5 show the results for $\beta$ Pictoris and HR 10. Concerning $\beta\:$ Pictoris, apart from the variations discussed by Lecavelier et al. (1995a), no other variations were significantly detected in these data, except a slight increase of the star brightness from 1995 (Fig. 4). This increase is marginal and at the limit of the detectability on this survey with a variation slope of $(-5.1 \pm 2.5(1\sigma))~10^{-6}$  mag per day. This is however noteworthy because this can be exactly superimposed on the variation detected with the Stömgren survey which will be discussed in Sect. 4.3.

For HR 10, we have already detected a short-term variation in January 1990 (Lecavelier et al. 1995a, and Fig. 5). The new data obtained since 1995 also show an apparent group of measurements around December 25, 1996 (Julian Date JD 2450443) during which the star appeared fainter than normally. But a careful look reveals that this group consists of three isolated measurements of lower quality, separated by good quality measurements at the normal brightness level (Fig. 5). We thus conclude that this apparent variation is not significant.

TY CrA is well known as an high amplitude eclipsing binary. We observed it to search for long term variations. Except one very short term light decrease related to an eclipse, no significant long term variations were observed.

68 Ophiuchi has been removed from the SAT program. This star was included in the Geneva program because of the suspected presence of circumstellar dust which was dismissed by subsequent observations (Vidal-Madjar et al. 1995; Waters et al. 1995).


  \begin{figure}
\includegraphics[width=8cm,clip]{bp_gen_2.ps}\end{figure} Figure 4: Plot of the Geneva V magnitude for $\beta\:$ Pictoris from 1976 to 1997 (middle panel). The filled circles represent the good quality data (weight $\ge 3$) and crosses represent the lower quality measurements (weight <3). This plot show the new data obtained from 1995 to 1997 in addition to the Fig. 1 of Lecavelier et al. (1995a). The upper panel show the rapid variation observed around November, 10 1981. The bottom panel show the long-term variation marginally detected here from 1995 and the corresponding linear fit to the data


  \begin{figure}
\includegraphics[width=8cm,clip]{hr10_1.ps}\end{figure} Figure 5: Plot of the Geneva V magnitude measured on HR 10 from 1990 to 1997 (middle panel). The two other panels show the times when the star appeared fainter than normal in January 1990 (top panel) and December 1996 (bottom panel). The filled circles represent the good quality data (weight $\ge 3$) and crosses represent the lower quality measurements (weight <3). It is clear that the variation in 1990 was real (and observed in other bands not shown here) while the apparent variation in late 1996 is only due to 3 lower quality measurements separated by good quality measurements at normal brightness

4.2 HD 38392 periodic variations

We were surprised to see in our search for short term variations that one of our stars, HD 38392 was presenting a surprisingly high magnitude of fluctuation, its sigma being at least twice the one observed in the case of the other stars of our program. Because, as already said, its long term variation was not presenting any trend, we tried to find out the cause of such a behaviour.

The magnitude difference between the stars HD 38392 and HD 38393 is shown for illustration as a function of the time (Fig. 6, top panel). Differences with the two other comparison stars definitely show that variations are due to HD 38392 and not to HD 38393.

From the observation of these data it appears that a periodic variation is present in the data. This was searched for by standard Fourier analysis and a very clear period of 21.4 days showed up immediately. Folding the data in phase, it shows a clear stellar variation of about 0.02 magnitude (Fig. 6, bottom panel), compatible with a fluctuation due to stellar activity with a rotation period of the star of the order of 21 days.

Note that this period was not previously known but only indirectly inferred (e.g. Soderblom et al. 1991, infer 12.8 days; Stepien & Geyer 1995, infer 20 days) from the correlation between the emission flux in the h and k lines of Mg II and the rotation period. Here we directly observed the stellar rotation period if our interpretation is correct.


  \begin{figure}
\includegraphics[width=8cm,clip]{phase.ps}\end{figure} Figure 6: Top panel presents the magnitude difference between HD 38392 and HD 38393 from the data gathered during three campaigns between 1995 and 1998. The slope of the linear fit to the data is very close to zero (solid line, top panel). There is no long-term variation for these two stars. Bottom panel shows the periodic variation detected for HD 38392 with a period of 21.4 days corresponding to its rotational period (Stepien & Geyer 1995)

   
4.3 $\beta\:$ Pictoris long term variations

In the $\beta\:$ Pictoris case, the calculated slope $b_{\beta{\rm Pic}}$ is significantly different from 0. The probability to have a so large slope $b_{\beta{\rm Pic}}$only by chance and due to the statistical noise is $P(b>b_{\beta{\rm Pic}}) \ll 10^{-7} $. The fits to the $\beta\:$ Pictoris data (solid lines in Figs. 2 and 3) start from JD 2450100 because before this date, data were obtained only with HD 35580 as comparison star.

The fit gives a slope $(-6.55 \pm 1.18(1\sigma))~10^{-6}$ mag per day if we use HD 40200 as comparison star, and a slope $(-5.56 \pm 0.75(1\sigma))~10^{-6}$ mag per day with HD 35580 as comparison star. Assuming that the distributions are Gaussian, we can combine the two slopes and we find that the variation of the $\beta\:$ Pictoris magnitude was $(-5.84 \pm 0.64(1\sigma))~10^{-6}$ mag per day. This corresponds to a variation of -2.1 10-3 mag per year.

As a conclusion, the brightness of $\beta\:$ Pictoris significantly increased from JD 2450100 (January 17, 1996) to JD 2450876 (March 3, 1998) by about 4.5 10-3 magnitude.


next previous
Up: A three-year Strömgren photometric stars

Copyright The European Southern Observatory (ESO)