Confronting sparse <i>Gaia</i> DR3 photometry with TESS for a sample of around 60 000 OBAF-type pulsators

Context. The Gaia mission has delivered hundreds of thousands of variable star light curves in multiple wavelengths. Recent work demonstrates that these light curves can be used to identify (non-)radial pulsations in OBAF-type stars, despite their irregular cadence and low light curve precision, of the order of a few millimagnitudes. With the considerably more precise TESS photometry, we revisited these candidate pulsators to conclusively ascertain the nature of their variability. Aims. We seek to re-classify the Gaia light curves with the first two years of TESS photometry for a sample of 58 970 p - and g -mode pulsators, encompassing γ Dor, δ Scuti, slowly pulsating B, and β Cep variables. From the TESS data, we seek to assess the quality of Gaia ’s classification of non-radial pulsators, which is based on sparse, years-long light curves of millimagnitude precision. We also supply four new catalogues containing the confirmed pulsators, along with their dominant and secondary pulsation frequencies, the number of independent mode frequencies, and a ranking according to their usefulness for future asteroseismic ensemble analysis. Methods. We first analysed the TESS light curves independent of their Gaia classification by pre-whitening all dominant pulsation modes down to a 1% false alarm probability. Using this, in combination with a feature-based random forest classifier, we identified different variability types across the sample. Results. We find that the Gaia photometry is exceptionally accurate for detecting the dominant and secondary frequencies, reaching approximately 80% accuracy in frequency for p - and g -mode pulsators. The majority of Gaia classifications are consistent with the classifications from the TESS data, illustrating the power of the low-cadence Gaia photometry for pulsation studies. We find that the sample of g -mode pulsators forms a continuous group of variable stars along the main sequence across B, A, and F spectral types, implying that the mode excitation mechanisms for all these pulsators need to be updated with improved physics. Finally, we provide a rank-ordered table of pulsators according to their asteroseismic potential for follow-up studies, based on the number of sectors they have been observed in, their classification probability, and the number of independent modes found in the TESS light curves from the nominal mission. Conclusions. Our catalogue offers a major increase in the number of confirmed g -mode pulsators with an identified dominant mode suitable for follow-up TESS ensemble asteroseismology of such stars.