Surveying flare rates and properties in a new wavelength regime with GBTDS
Program ID 19104
Science Category Stellar Populations
Program Type Analysis
Category Medium
Principal Investigator Ward Howard
PI Institution University of Colorado, Boulder
Co-Investigators
  • Hank Corbett (University of North Carolina, Chapel Hill)
  • Adina Feinstein (Michigan State University)
  • Zachary Claytor (Space Telescope Science Institute / STScI)
  • Guadalupe Tovar Mendoza (Johns Hopkins University)
  • Meredith MacGregor (Johns Hopkins University)
Abstract Only several dozen flare light curves have been recorded from a handful of stars at infrared wavelengths, preventing any unified understanding of flare mechanisms probed by this wavelength regime or comparison to better-studied wavelength regimes. The stochastic occurrence and rarity of large flares, short timescales, and low fluxes at IR wavelengths make Roman’s GBTDS the only current or planned facility capable of obtaining robust statistics on IR flare rates and properties as functions of stellar mass, age, rotation, and binarity. We propose a deep learning-assisted flare survey across all 25M sources of F146<27 in the first three high cadence seasons of the GBTDS to detect 398,000 flares, including 64,000 from sources with detectable stellar rotation periods to infer approximate gyrochronological ages. Our survey will (1) determine whether infrared flare occurrence rates follow a single or broken-power law distribution, (2) identify differences in the physical processes underlying flare emission at infrared compared to optical wavelengths, and (3) compare flare rates and morphologies at 99.7% confidence across a 2D grid of stellar masses and ages from 0.1-1.6 solar masses and 10-10,000 Myr to assess the impacts of spin down on IR flaring. Our deep learning flare detection software, flare light curves, and catalog will be made publicly available to support flare analysis in the wider Roman community.