Subscribe to the gold package and get unlimited access to Shamra Academy
Register a new userThe Cannon: A data-driven approach to stellar label determination
192
0
0.0
(
0
)
Ask ChatGPT about the research
No Arabic abstract
New spectroscopic surveys offer the promise of consistent stellar parameters and abundances (stellar labels) for hundreds of thousands of stars in the Milky Way: this poses a formidable spectral modeling challenge. In many cases, there is a sub-set of reference objects for which the stellar labels are known with high(er) fidelity. We take advantage of this with The Cannon, a new data-driven approach for determining stellar labels from spectroscopic data. The Cannon learns from the known labels of reference stars how the continuum-normalized spectra depend on these labels by fitting a flexible model at each wavelength; then, The Cannon uses this model to derive labels for the remaining survey stars. We illustrate The Cannon by training the model on only 542 stars in 19 clusters as reference objects, with Teff, log g and [Fe/H] as the labels, and then applying it to the spectra of 56,000 stars from APOGEE DR10. The Cannon is very accurate. Its stellar labels compare well to the stars for which APOGEE pipeline (ASPCAP) labels are provided in DR10, with rms differences that are basically identical to the stated ASPCAP uncertainties. Beyond the reference labels, The Cannon makes no use of stellar models nor any line-list, but needs a set of reference objects that span label-space. The Cannon performs well at lower signal-to-noise, as it delivers comparably good labels even at one ninth the APOGEE observing time. We discuss the limitations of The Cannon and its future potential, particularly, to bring different spectroscopic surveys onto a consistent scale of stellar labels.
rate research
Read More
Upcoming million-star spectroscopic surveys have the potential to revolutionize our view of the formation and chemical evolution of the Milky Way. Realizing this potential requires automated approaches to optimize estimates of stellar properties, such as chemical element abundances, from the spectra. The volume and quality of the observations strongly motivate that these approaches should be data-driven. With this in mind, we introduce SSSpaNG: a data-driven non-Gaussian Process model of stellar spectra. We demonstrate the capabilities of SSSpaNG using a sample of APOGEE red clump stars, whose model parameters we infer via Gibbs sampling. Pooling information between stars to infer their covariance, we permit clear identification of the correlations between spectral pixels. Harnessing these correlations, we infer the true spectrum of each star, inpainting missing regions and denoising by a factor of at least 2 for stars with signal-to-noise of ~20. As we marginalize over the covariance matrix of the spectra, the effective prior on these true spectra is non-Gaussian and sparsifying, favouring typically small but occasionally large excursions from the mean. The high-fidelity inferred spectra produced will enable improved elemental abundance measurements for individual stars. Our model also allows us to quantify the information gained by observing portions of a stars spectrum, and thereby define the most mutually informative spectral regions. Using 25 windows centred on elemental absorption lines, we demonstrate that the iron-peak and alpha-process elements are particularly mutually informative for these spectra, and that the majority of information about a target window is contained in the 10-or-so most informative windows. Such mutual-information estimates have the potential to inform models of nucleosynthetic yields and the design of future observations.
The RAdial Velocity Experiment (RAVE) is a spectroscopic survey of the Milky Way. We use the subsample of spectra with spectroscopically determined values of stellar parameters to determine the distances to these stars. The list currently contains 235,064 high quality spectra which show no peculiarities and belong to 210,872 different stars. The numbers will grow as the RAVE survey progresses. The public version of the catalog will be made available through the CDS services along with the ongoing RAVE public data releases. The distances are determined with a method based on the work by Breddels et al.~(2010). Here we assume that the star undergoes a standard stellar evolution and that its spectrum shows no peculiarities. The refinements include: the use of either of the three isochrone sets, a better account of the stellar ages and masses, use of more realistic errors of stellar parameter values, and application to a larger dataset. The derived distances of both dwarfs and giants match within ~21% to the astrometric distances of Hipparcos stars and to the distances of observed members of open and globular clusters. Multiple observations of a fraction of RAVE stars show that repeatability of the derived distances is even better, with half of the objects showing a distance scatter of simlt 11%. RAVE dwarfs are ~300 pc from the Sun, and giants are at distances of 1 to 2 kpc, and up to 10 kpc. This places the RAVE dataset between the more local Geneva-Copenhagen survey and the more distant and fainter SDSS sample. As such it is ideal to address some of the fundamental questions of Galactic structure and evolution in the pre-Gaia era. Individual applications are left to separate papers, here we show that the full 6-dimensional information on position and velocity is accurate enough to discuss the vertical structure and kinematic properties of the thin and thick disks.
To accurately interpret the observed properties of exoplanets, it is necessary to first obtain a detailed understanding of host star properties. However, physical models that analyze stellar properties on a per-star basis can become computationally intractable for sufficiently large samples. Furthermore, these models are limited by the wavelength coverage of available spectra. We combine previously derived spectral properties from the Spectroscopic Properties of Cool Stars (SPOCS) catalog (Brewer et al. 2016) with generative modeling using The Cannon to produce a model capable of deriving stellar parameters ($log g$, $T_{mathrm{eff}}$, and $vsin i$) and 15 elemental abundances (C, N, O, Na, Mg, Al, Si, Ca, Ti, V, Cr, Mn, Fe, Ni, and Y) for stellar spectra observed with Keck Observatorys High Resolution Echelle Spectrometer (HIRES). We demonstrate the high accuracy and precision of our model, which takes just $sim$3 seconds to classify each star, through cross-validation with pre-labeled spectra from the SPOCS sample. Our trained model, which takes continuum-normalized template spectra as its inputs, is publicly available at https://github.com/malenarice/keckspec. Finally, we interpolate our spectra and employ the same modeling scheme to recover labels for 477 stars using archival stellar spectra obtained prior to Kecks 2004 detector upgrade, demonstrating that our interpolated model can successfully predict stellar labels for different spectrographs that have (1) sufficiently similar systematics and (2) a wavelength range that substantially overlaps with that of the post-2004 HIRES spectra.
The orbits of binary stars and planets, particularly eccentricities and inclinations, encode the angular momentum within these systems. Within stellar multiple systems, the magnitude and (mis)alignment of angular momentum vectors among stars, disks, and planets probes the complex dynamical processes guiding their formation and evolution. The accuracy of the textit{Gaia} catalog can be exploited to enable comparison of binary orbits with known planet or disk inclinations without costly long-term astrometric campaigns. We show that textit{Gaia} astrometry can place meaningful limits on orbital elements in cases with reliable astrometry, and discuss metrics for assessing the reliability of textit{Gaia} DR2 solutions for orbit fitting. We demonstrate our method by determining orbital elements for three systems (DS Tuc AB, GK/GI Tau, and Kepler-25/KOI-1803) using textit{Gaia} astrometry alone. We show that DS Tuc ABs orbit is nearly aligned with the orbit of DS Tuc Ab, GK/GI Taus orbit might be misaligned with their respective protoplanetary disks, and the Kepler-25/KOI-1803 orbit is not aligned with either components transiting planetary system. We also demonstrate cases where textit{Gaia} astrometry alone fails to provide useful constraints on orbital elements. To enable broader application of this technique, we introduce the python tool texttt{lofti_gaiaDR2} to allow users to easily determine orbital element posteriors.
Log in to be able to interact and post comments
comments
Fetching comments
Sign in to be able to follow your search criteria
