skip to main content

gaia data release 3 documentation

10.8 Long period variables

10.8.3 Calibration models

The periods of the G light curves of the LPV candidates are searched using a standard Least Squares method. The first period is retained for publication for each source. A Fourier series is then fit to the light curve with that period to get the amplitude of the model.

The classification between C-rich and O-rich LPVs is based on the RP spectra time series, taking advantage of the distinctive features found in the spectra of each type. The spectra of O-rich red giants display numerous molecular absorption bands, mainly due to TiO and related oxides. The absorption features in the spectra of C-rich red giants, on the other hand, are fewer but wider, mainly associated with CH and C2 molecules. These features are visible in low-resolution RP spectra, and are used to distinguish between C-rich and O-rich stars in the Gaia catalogue. The distinction proceeds with the computation of the pseudo-wavelength separation of the two highest peaks in the RP spectra, which is taken as a proxy for the distance and thus location of the molecular absorption bands characteristic for either C-rich or O-rich chemistry. An example of the RP spectra of each type and of the classification results have been shown in the Gaia image of the week of 15 November 2018. The details of the procedure are given in Section 10.8.4.