MWS SPDist Catalog¶
Overview¶
This catalog provides spectrophotometric distances in the form of absolute magnitude in the Gaia G-band for all stars observed by the DESI MWS, using a data-driven approach similar to Thomas & Battaglia (2022. Distances are estimated using a fully connected multi-layer perceptron (MLP) model using Gaia DR3 photometry (Gaia Collaboration et al. 2023), spectroscopic metallicity from the SP pipeline, and atmospheric parameters (effective temperature and surface gravity) derived from a combination of values from both the RV and SP pipelines (Koposov et al. 2024) as inputs. The table here is row-matched to DESI MWS Iron release. Note that SPdist stands for Spectro-photometric distances and is not related to the SP pipeline used in the MWS Value Added Catalog.
The DESI Milky Way Survey provides two stellar distance Value Added Catalogs: SPdist (this catalog), and SpecDis. The major difference between these two catalogs lies in the method used to derive the distances. SpecDis predicts distances directly from stellar spectra, while SPdist uses the stellar parameters provided by the DESI Milky Way Survey pipelines.
Data Access¶
Data URL: https://data.desi.lbl.gov/public/dr1/vac/dr1/mws-spdist
NERSC access:
/global/cfs/cdirs/desi/public/dr1/vac/dr1/mws-spdist
Documentation¶
Files¶
mws-iron-SPDist-240613.fits: Catalog of distances for DR1.
Data Model¶
EXTNAME=SPDIST
| Name | Type | Units | Description |
|---|---|---|---|
| CAT_ID | int64 | - | Entry ID, matching the main DESI MWS Iron catalog. |
| REF_ID | int64 | - | Reference ID from the DESI MWS Iron catalog. |
| TARGETID | int64 | - | ID. of the target |
| SOURCE_ID | int64 | - | Gaia DR3 identifier |
| FLAG_GOOD | bool | - | True if the TEFF and LOGG parameters estimated by SP and RVS pipelines are within 3-sigma of the median distribution |
| TEFF_COR | float64 | K | Average temperature of the SP and RVS pipeline, calibrated on the RVS base |
| LOGG_COR | float64 | - | Average SP and RVS logg |
| E_Gmag | float64 | mag | Gaia G band uncertainties |
| E_BPmag | float64 | mag | Gaia BP band uncertainties |
| E_RPmag | float64 | mag | Gaia RP band uncertainties |
| G0 | float64 | mag | derreded Gaia G band photometry (using i the Schlegel et al. 1998 extinction map) |
| BP0 | float64 | mag | derreded Gaia BP band photometry |
| RP0 | float64 | mag | derreded Gaia RP band photometry |
| MG_PEAK | float64 | mag | Absolute Magnitude using the peak value of TEFF (RV & SP), LOGG(RV&SP), FEH (SP), BP0 and RP0 |
| MG_5 | float64 | mag | 5th percentile of the absolute magnitude distribution after the propagation of the uncertainties on the TEFF (RV & SP), LOGG(RV&SP), FEH (SP), BP and RP |
| MG_16 | float64 | mag | 16th percentile of the absolute magnitude distribution after the propagation of the uncertainties on the TEFF (RV & SP), LOGG(RV&SP), FEH (SP), BP and RP |
| MG_50 | float64 | mag | median of the absolute magnitude distribution after the propagation of the uncertainties on the TEFF (RV & SP), LOGG(RV&SP), FEH (SP), BP and RP; Recommanded value to use |
| MG_84 | float64 | mag | 84th percentile of the absolute magnitude distribution after the propagation of the uncertainties on the TEFF (RV & SP), LOGG(RV&SP), FEH (SP), BP and RP |
| MG_95 | float64 | mag | 95th percentile of the absolute magnitude distribution after the propagation of the uncertainties on the TEFF (RV & SP), LOGG(RV&SP), FEH (SP), BP and RP |
| QFLAG_INPUT | bool | - | True if the parameters space are located inside the region define by the training sample |
| QFLAG_COMMENTS | str | - | Say the reason why the Qflag_input is False |
Recommended Use¶
The presented catalog contains stars flagged with suspicious parameters. Some of these stars are located outside the training sample’s parameter space, leading to extrapolated values. For this reason, we recommend the following:
- Some stars exhibit significant discrepancies in spectroscopic parameters derived by the RV and SP pipelines, likely resulting in unreliable distance estimates. These suspicious stars are flagged with the
FLAG_GOODparameter. We recommend using this flag along with the quality flags suggested for the RV and SP pipelines (Koposov et al. 2024). - For distance modulus computations, we recommend using the median of the absolute magnitude distribution (
MG_50), while uncertainties can be derived from the 16th and 84th percentiles of the distribution. These uncertainties should be combined in quadrature with the intrinsic precision of the method (0.167 mag, 8% relative distance precision) to obtain the total uncertainty.
Contact¶
Contact Guillaume Thomas for questions about this catalog.