A Distance Measurement to M33 Using Optical Photometry of Mira Variables

TL;DR

This paper refines the distance to M33 by analyzing optical light curves of Mira variables over 18 years, employing machine learning for classification, and deriving a precise distance modulus consistent with previous estimates.

Contribution

It introduces a systematic method combining long-term optical photometry and machine learning to improve Mira period determination and distance measurement to M33.

Findings

Optical light curves are essential for accurate Mira period determination.

Machine learning effectively classifies Miras into O-rich and C-rich types.

Derived distance modulus to M33 is 24.67 ± 0.06 mag, aligning with literature.

Abstract

We present a systematic analysis to determine and improve the pulsation periods of 1637 known long-period Mira variables in M33 using $gri$-band light curves spanning $\sim18$~years from several surveys, including M33 variability survey, Panoramic Survey Telescope and Rapid Response System, Palomar Transient Factory (PTF), intermediate PTF, and Zwicky Transient Facility. Based on these collections of light curves, we found that optical band light curves that are as complete as possible are crucial to determine the periods of distant Miras. We demonstrated that the machine learning techniques can be used to classify Miras into O-rich and C-rich based on the $(J-K_s)$ period--color plane. Finally, We derived the distance modulus to M33 using O-rich Miras at maximum light together with our improved periods as $24.67 \pm 0.06$~mag, which is in good agreement with the recommended value given in the literature.