Best¶
The Best pipeline (internally MWMBest) selects the single best set of stellar parameters
and auxiliary measurements for each source observed by the Milky Way Mapper (MWM), drawing
from across all available analysis pipelines.
What it does¶
Best produces one row per source that consolidates:
Stellar parameters (Teff, log g, metallicity, abundances)
Radial velocities from multiple methods
Chemical abundances (up to 20+ elements)
White dwarf classifications (from Snow White)
M dwarf spectral types (from MDwarfType)
Spectrum metadata (SNR, MJD range, number of visits)
How it works¶
The Best pipeline queries results from multiple analysis pipelines in a priority-ordered sequence. For each source, it takes the result from the first pipeline in the sequence that has a valid measurement. Once a source has been assigned results, it is excluded from subsequent queries.
Pipeline priority order¶
Snow White (white dwarf classifications) – for sources in the
mwm_wdprogramBOSSNet (hot stars) – for sources in the
mwm_obprogram, with no result flagsASPCAP (hot stars) – for sources in the
mwm_obprogramAPOGEENet (YSOs) – for sources in the
mwm_ysoprogram, with no result flagsBOSSNet (YSOs) – for sources in the
mwm_ysoprogram, with no result flagsSLAM (M dwarfs) – for sources with no result flags
ASPCAP (general) – for sources not flagged as bad
AstroNN (general) – for sources with no result flags
APOGEENet (general) – for sources with no result flags
BOSSNet (general) – for sources with no result flags
APOGEENet (any remaining) – no quality cuts
BOSSNet (any remaining) – no quality cuts
This priority order reflects the science strategy: specialized pipelines are preferred for their target populations (e.g., Snow White for white dwarfs, SLAM for M dwarfs), and higher-quality results are preferred over lower-quality ones.
Spectrum metadata¶
Each Best result also carries metadata from the underlying spectrum:
APOGEE spectra: release, apred, apstar, telescope, field, fiber information, Doppler RV parameters, cross-correlation RV, visit statistics
BOSS spectra: release, run2d, telescope, XCSAO parameters, zwarning flags, visit statistics
Output fields¶
Radial velocity¶
Field |
Description |
|---|---|
|
Best radial velocity (km/s) |
|
Uncertainty on radial velocity |
|
Standard deviation of visit radial velocities |
|
Median per-visit RV uncertainty |
Radial velocity (method-specific)¶
Field |
Description |
|---|---|
|
XCSAO template parameters |
|
XCSAO cross-correlation R-value |
|
Doppler template parameters |
|
Doppler fit reduced chi-squared |
|
Cross-correlation RV results |
|
BOSSNet radial velocity |
Stellar parameters¶
Field |
Description |
|---|---|
|
Effective temperature (K) |
|
Surface gravity (log cm/s^2) |
|
Microturbulent velocity (km/s) |
|
Projected rotational velocity (km/s) |
|
Overall metallicity [M/H] (dex) |
|
Alpha-element abundance [alpha/M] (dex) |
|
Atmospheric carbon abundance [C/M] (dex) |
|
Atmospheric nitrogen abundance [N/M] (dex) |
Chemical abundances¶
Individual element abundances are provided for: Al, C, C_1, Ca, Ce, Co, Cr, Cu, Fe, K, Mg, Mn, N, Na, Nd, Ni, O, P, S, Si, Ti, Ti_2, V, and C_12_13 (carbon isotope ratio).
Each element X has four fields:
X_h: Abundance [X/H] (dex)e_X_h: UncertaintyX_h_flags: Quality flagsX_h_rchi2: Reduced chi-squared for the abundance window fit
White dwarf classifications¶
Field |
Description |
|---|---|
|
White dwarf classification string |
|
Probabilities for each white dwarf subtype |
M dwarf classifications¶
Field |
Description |
|---|---|
|
M dwarf spectral type |
|
Numerical sub-type |
Observing metadata¶
Field |
Description |
|---|---|
|
Signal-to-noise ratio |
|
MJD range of visits |
|
Number of visits |
|
Number of good visits |
|
Number of good radial velocities |
|
Telescope used |
|
Data release and file type |
Key caveats¶
The Best result for a given source depends on which pipelines ran successfully and the priority order. The “best” designation reflects the operational strategy, not necessarily the most accurate measurement for a particular science case.
A single source has at most one Best entry (unique on
source_pkandv_astra). If a source was observed by both APOGEE and BOSS, the Best result comes from whichever pipeline ranks higher in the priority list.Fields that are not provided by the selected pipeline will be null. For example, chemical abundances are only populated when ASPCAP or AstroNN is the selected pipeline.
The
pipeline_flagsfield is an amalgamated bitfield that may combine flags from the selected source pipeline.Users interested in results from a specific pipeline should query that pipeline’s table directly rather than relying on Best.