FUMES III: Ultraviolet and Optical Variability of M Dwarf Chromospheres

We obtained ultraviolet and optical spectra for 9 M dwarfs across a range of rotation periods to determine whether they showed stochastic intrinsic variability distinguishable from flares. The ultraviolet spectra were observed during the Far Ultraviolet M Dwarf Evolution Survey Hubble Space Telescope program using the Space Telescope Imaging Spectrograph. The optical observations were taken from the Apache Point Observatory 3.5-meter telescope using the Dual Imaging Spectrograph and from the Gemini South Observatory using the Gemini Multi-Object Spectrograph. We used the optical spectra to measure multiple chromospheric lines: the Balmer series from H$\alpha$ to H$10$ and the Ca II H and K lines. We find that after excising flares, these lines vary on the order of $1-20\%$ at minute-cadence over the course of an hour. The absolute amplitude of variability was greater for the faster rotating M dwarfs in our sample. Among the 5 stars for which measured the weaker Balmer lines, we note a tentative trend that the fractional amplitude of the variability increases for higher order Balmer lines. We measured the integrated flux of multiple ultraviolet emission features formed in the transition region: the N V, Si IV, and C IV resonance line doublets, and the C II and He II multiplets. The signal-to-noise (S/N) ratio of the UV data was too low for us to detect non-flare variability at the same scale and time cadence as the optical. We consider multiple mechanisms for the observed stochastic variability and propose both observational and theoretical avenues of investigation to determine the physical causes of intrinsic variability in the chromospheres of M dwarfs.

All code associated with the analysis and plots for this paper are included in .py scripts. All raw data for the optical spectra are its files compressed into a tar.gzip archive while all reduced spectra are in fits files. All tables are provided in the astropy ASCII text ecsv file format.

Optical Reduction Code: The reduction code is divided into two folders, one labelled "pydis" and another labelled "pygemini".

Equivalent Width Measurements: The equivalent widths are measured using a combination of two tables for each exposure: one ending with the suffix "ew_windows.ecsv" that lists the boundaries of the blue and red continua windows and the continuum flux density value, and another ending with the suffix "ew_lines.ecsv" that lists the boundaries of the wavelength window, the integrated line flux with its error, and the equivalent width with its error. These tables are in a subdirectory named "fit_tables".

spectralPhoton: The version of spectralPhoton code and the scripts we use to split the Hubble x1d spectra are in a directory named "uv".

Line-fitting Tables: The parameter values and associated errors are recorded in tables structured similarly to the equivalent width tables, ending with suffixes "windows.ecsv" and "lines.ecsv".

Time Series Tables: The equivalent widths and integrated line fluxes are collated into time series tables in the subdirectory "time_series". Entries with cosmic ray hits in the middle of the line or other spectral defects have been commented out using a # symbol.

Posterior Distributions: All posterior samples and their log-likelihood values are recorded in numpy binary .npy files that can be read using the numpy.load() function. These files are divided into two subdirectories named ``abs"  and ``frac" for the absolute and fractional flux fits respectively.