MARVELS Target Selection

Between October 2008 and July 2012 MARVELS made 1565 observations of 95 fields collecting multi-epoch data for 5700 stars, aiming at 60 stars per target field (although not all of these fields were observed with sufficient quality to appear in the final archives of DR11 and DR12 reduced RV curves). This page briefly describes how the target stars have been selected.

Due to technical and administrative changes in January 2011 — change of fibers, joined observation with APOGEE — the observation is cut into 2 different phases: before and after January 2011, further referred as initial and final target selection. The change in observation strategy was accompanied with a change in target selection.

Initial target selection

Based on a modified Guide Star Catalog 2.3 which included proper motions for the northern hemisphere that has been matched to 2MASS, Tycho-2 and Hipparcos, the up to 1000 brightest stars matching the following criteria have been selected:

  • 7.6 ≤ V ≤ 13.0
  • must have known proper motions
  • must have JHK colors with J-K ≥ 0.29
  • no star brighter than V = 9.0 within 5 arcsec
  • no other target star closer than 62 arcsec
  • must stay on plate for at least 2 years

Selected stars were observed with the BOSS spectrograph during astronomical twilight (This is known as Preselection data). The spectra were processed through a modified version of the SSPP pipeline the following stellar parameters were extracted:

  • effective temperature (Teff)
  • surface gravity (log(g)) (See Footnote)
  • metallicity ([Fe/H])

Once these values were known, the stars were split into a bright (V ≤ 9.0) and a faint sample (up to V = 13.0). Using SIMBAD the bright stars were manually checked and accepted if

  • (spectral type is between F and early K and log(g) ≥ 3.0) or (spectral type is between mid G and early K and log(g) < 3.0)
  • star is not a known variable
  • star is not a multiple with a companion closer than 5 arcsec
  • star is not a known exoplanet host
  • star has no anomalies making it unlikely to find substellar companions

The remaining stars from the bright sample have been rejoined and the final 100 stars for observation with MARVELS have been selected using the following criteria:

  • keep the 6 brightest stars with log(g) ≤ 3.0
  • limit F8 to G0 mains sequence stars to no more than 40% of the whole sample
  • keep the brightest 100 stars

We keep the brightest 100 stars, so that we have a “reserve” in case of conflicts with guide stars or other technical details force us to go down the list of available stars in order of brightness. This results in a plate that contains the 60 brightest stars that were not removed due to conflicts.

Final target selection

From January 2011 on MARVELS and APOGEE observed the same target field at the same time, including many southern fields for which the modified Guide Star Catalog had no proper motions. For these fields we matched Nomad, UCAC3, Tycho-2 and Hipparcos to a MARVELS Input Catalog.

At the same time, we replaced the observations by BOSS (Preselection) and used the Infrared Flux Method for computing the effective temperature. We further used a reduced proper motion cut to distinguish between dwarfs and giants. We defined a dwarf by log(g) ≥ 4.1 and a giant by log(g) ≤ 3.5. Hipparcos and the Kepler Input Catalog were used as benchmarks to define a cut in reduced proper motion space that separates these two populations. We found that this cut distinguishes giants from dwarfs with a false positive rate of 8 – 10% between both groups for the spectral types targeted By MARVELS. Using the reduced proper motion cut sub-giants (3.5 < log(g) < 4.1) could not be isolated from dwarfs, but this is not a concern because non-giant stars are valid MARVELS targets.

While only the 6 brightest giants were manually checked as described above, all other criteria for brightness and spectral types respectively effective temperatures were kept.


Modified SSPP and MARVELS Preselection Data

Later analysis of the surface gravities showed the results from the modified SSPP pipeline were not reliable. MARVELS was designed to have 10% of the targets be giants, but the unreliable surface gravities lead to 30% of the targets being giants in the year 1 + 2 data.

Teff provided for year 1 + 2 vs. year 3 + 4 data

Since preselection data were only available for years 1 + 2, SSPP temperatures could not be assigned to stars monitored in years 3 + 4. So, temperatures provided for years 3 + 4 are from the Casagrande IRFM methodology using V-K or J-K, depending on which is available, or, if both are available, then the average of the two results. Temperatures provided for years 1 + 2 are from SSPP.