Detecting short period variable stars with Gaia

We analyzed the frequency domain of time series of simulated ZZ Ceti light-curves to investigate the detectability and period recovery performance of short period variables (periods < 2 hours) for the Gaia mission. In our analysis, first we used a non-linear ZZ Ceti light-curves simulator code to simulate the variability of ZZ Ceti stars (we assumed stationary power spectra over five years). Second we used the Gaia nominal scanning law and the expected photometric precision of Gaia to simulate ZZ Ceti time series with Gaia's time sampling and photometric errors. Then we performed a Fourier analysis of these simulated time series. We found that a correct period can be recovered in ~65% of the cases if we consider Gaia per CCD time series of a G ~ 18 magnitude multiperiodic ZZ Ceti star with 5%-10% light-curve variation. In the pre-whitened power spectrum a second correct period was also recovered in ~26% of the cases.