Abstract.The Gaia mission characteristics permit to estimate its impact on the variable star research. During its five-year operation Gaia will produce about 100 photometric measurements in each passband of the Sloan system for about 1 billion stars. Thanks to our knowledge of the Hipparcos data and the similarities between these two missions, we can forecast that Gaia will detect tens of millions of new variable stars.Key words: stars: variables, photometry -orbiting observatories: Gaia
INTRODUCTIONThe first step of variable star analysis is to study the time series of individual measurements per object for obtaining and to reduce them to mean magnitudes, light curves, intrinsic dispersions, noises, periods, epochs, light curves, etc. The ability to obtain this information depends on the signal parameters (amplitude, shape) and the measurement and reduction procedures in the satellite (time sampling, number of measurements, noise). The accurate estimation of the noise and the ability to identify outlying values are two crucial prerequisites for variable star analysis. Therefore, precise calibration procedures are needed, e.g. the control of star extraction in crowded fields, controlled ageing effects, etc.
ESTIMATION OF THE NUMBER OF VARIABLE STARSWe binned the star population measured by Gaia, and applied the fraction of variable stars found in the Hipparcos analysis for the same population and the same magnitude precision. The model inputs are the following:• The errors of individual measurements. In our estimation we adopted their magnitude dependence from H0g et al. (1998) for the G magnitude and an error of 0.01 mag for V < 17). The saturation affects only a small number of bright stars and is not taken into account.• The star population.We have used the starcount model (iV(My), N(my) as a function of V-I) given by Torra et al. (1999). The total number of stars with V < 20 mag is 912 millions.• The variability detection threshold for the foreseen Gaia time sampling and noise distribution as a function of magnitude (established from Hipparcos results. With this model, we estimate 26 million variable stars, among them 7 million periodic variables.Some comments can be added: • The Gaia star sample will differ from the Hipparcos sample by the presence of more numerous red dwarf stars. Red dwarfs generally are variable with pseudo-periodic variations of several days (BY Draconis stars). A large fraction of them can be detected with Gaia.• Because the Hipparcos survey is magnitude limited, sample biases are present when estimating the fraction of variables (for example, in the cases of EB or EW eclipsing binaries).• The Gaia time sampling will be semi-regular and will introduce aliasing frequencies when doing period search. This sampling will lower the ability to detect certain periods. In the Hipparcos database, the performance of the period search algorithms usually dropped when periods were close to 15 days (cf. Eyer et al 1994).• GAIA will also produce time series for stars in the Magellanic Clouds and for...