Milky Way Mapper (MWM): Galactic Spectroscopy

• Milky Way Mapper is a survey-scale stellar spectroscopy program within SDSS-V, aimed at mapping the structure, chemistry, and dynamics of the Milky Way.
• It employs dual instruments—high-resolution APOGEE H-band and low-resolution BOSS optical spectroscopy—to study diverse stellar populations from red giants to OBAF stars.
• The DR19 release provided approximately 1.2 million infrared and 800,000 optical spectra, enabling detailed chemical cartography, precision open cluster calibration, and insights into Galactic evolution.

Searching arXiv for recent Milky Way Mapper papers to ground the article in published work. {"queries":[{"q":"all:\"Milky Way Mapper\" SDSS-V DR19", "max_results": 10},{"q":"ti:\"Milky Way Mapper\" OR abs:\"Milky Way Mapper\"", "max_results": 10}]} Here are the top arXiv matches I found:

1. "SDSS-V Milky Way Mapper (MWM): ASPCAP Stellar Parameters and Abundances in SDSS-V Data Release 19" — (Mészáros et al., 9 Jun 2025)
2. "The Nineteenth Data Release of the Sloan Digital Sky Survey" — (Collaboration et al., 9 Jul 2025)
3. "Reconstructing the Milky Way chemical map with Galactic Chemical Evolution tool OMEGA+ from SDSS-MWM" — (Hegedűs et al., 31 May 2025)
4. "The Open Cluster Chemical Abundances and Mapping Survey: VIII. Galactic Chemical Gradient and Azimuthal Analysis from SDSS/MWM DR19" — (Otto et al., 9 Jul 2025)
5. "A Comprehensive Study of Open Cluster Chemical Homogeneity using APOGEE and Milky Way Mapper Abundances" — (Sinha et al., 2024)
6. "Milky Way Mapper decoded abundances -- I. Shared disc enrichment patterns" — (Ness et al., 19 May 2026)
7. "Milky Way Mapper decoded abundances -- II: From patterns to paths" — (Ness et al., 20 May 2026)
8. "Zeta-Payne: a fully automated spectrum analysis algorithm for the Milky Way Mapper program of the SDSS-V survey" — (Straumit et al., 2022)
9. "The Eighteenth Data Release of the Sloan Digital Sky Surveys: Targeting and First Spectra from SDSS-V" — (Almeida et al., 2023)
10. "A Data-Driven M Dwarf Model and Detailed Abundances for ~17,000 M Dwarfs in SDSS-V" — (Behmard et al., 24 Jan 2025) Milky Way Mapper (MWM) is the Milky Way–focused stellar spectroscopy program within SDSS-V and one of the survey’s three core “Mapper” programs, alongside the Black Hole Mapper and the Local Volume Mapper. In the SDSS-V framework, MWM is designed to obtain optical and infrared stellar spectroscopy at survey scale in order to chart Galactic structure, chemistry, kinematics, stellar populations, dust, and time-domain stellar phenomena. DR18 established the targeting databases, cartons, and selection-function machinery for MWM, while DR19 became the first release with substantial public MWM spectra and products, including approximately 1.2 million near-infrared APOGEE spectra and 800,000 optical BOSS spectra, corresponding to about 390,000 unique stars with infrared data and 475,000 unique stars with optical data (Almeida et al., 2023, Collaboration et al., 9 Jul 2025).

1. Program definition, scientific scope, and nomenclature

Within SDSS-V, a “Mapper” is a survey-scale observing program built around a coherent science theme, shared targeting infrastructure, and centralized target-selection cartons. MWM is the Mapper devoted to the Milky Way itself. The DR18 and DR19 survey papers describe it as the SDSS-V component dedicated to optical + infrared stellar spectroscopy of millions of Milky Way stars, with goals that include mapping the structure and dynamics of the Galaxy, measuring stellar parameters and chemical abundances, studying stellar populations in the disk, bulge, halo, and star-forming regions, constraining dust/extinction, identifying special stellar classes and time-domain objects, and providing a foundation for Galactic archaeology and stellar astrophysics at scale (Almeida et al., 2023, Collaboration et al., 9 Jul 2025).

DR19 presents MWM as a program that observes stars across the Hertzsprung–Russell diagram in order to decode Galactic history, understand stellar systems architecture, and probe stellar physics. Its thirteen overarching programs are listed as Galactic Genesis, Magellanic Genesis, White Dwarfs, Solar Neighborhood Census, Halo, Young Stellar Objects (YSOs), OB Stars, Dust, Galactic eROSITA Sources, Binary Systems, Compact Binaries, Planet Hosts, and Asteroseismic Red Giants (Collaboration et al., 9 Jul 2025). This breadth is central to the identity of MWM: it is not restricted to a single tracer population such as luminous giants, even though red giants remain the dominant backbone for many Galactic archaeology applications.

A recurrent source of confusion is nomenclature. In SDSS-V, Milky Way Mapper is the formal program name. In DESI, by contrast, the corresponding stellar survey is the Milky Way Survey (MWS), a bright-time Galactic component of DESI. The DESI literature describes MWS as the foundation for what is often referred to as Milky Way mapping science, but it is a distinct survey with different instrumentation, footprint, and selection design (Prieto et al., 2020, Cooper et al., 2022). This distinction matters when comparing results, because SDSS-V MWM is built around APOGEE high-resolution near-infrared spectroscopy plus BOSS optical spectroscopy, whereas DESI MWS is a low-resolution optical survey.

2. Targeting architecture and survey execution

MWM’s targeting philosophy is organized through input catalogs, target cartons, target-selection generations, and released selection metadata. DR18 was fundamentally a targeting and survey-definition release for MWM rather than a release of the main MWM spectra themselves, and it made public the targeting databases, cartons, and recoverable selection functions needed for later statistical analyses (Almeida et al., 2023). The targeting ecosystem draws on broad catalog cross-matches and records why each object was selected, so that the survey remains scientifically interpretable after the fact.

In DR19, MWM targeting is described through the shared MOS target product, with selection proceeding through cross-matched catalogs, cartons, target selection generations, robotic assignment by robostrategy, and nightly execution by roboscheduler. The data release includes both early plate-program targeting and early FPS-era targeting. The plate era was a temporary but operationally important phase caused by COVID-related delays, and MWM plate targets were used to begin science while the fiber-positioning system was still being commissioned (Collaboration et al., 9 Jul 2025).

Concrete carton definitions illustrate how MWM balances simplicity and specialization. The mwm_gg_core Galactic Genesis carton is described as a simple color–magnitude cut targeting luminous cool giant stars with $H < 11$, $G-H > 3.5$ or Gaia non-detection, additional quality flags, and cadence $1 \times 33$ min. The mwm_rv_long-bplates carton selects stars previously observed at least three times with APOGEE, with APOGEE visits $\geq 6$, $H < 12.2$, specific APOGEE target flags, Gaia-based distance information, and cadence options $6 \times 67$ min or $12 \times 67$ min. The mwm_yso_s1 carton selects YSOs with disks through infrared excess criteria such as $W1-W2 > 0.25$, $W2-W3 > 0.50$, $W3-W4 > 1.50$, and $G-H > 3.5$0 mas, with cadence $G-H > 3.5$1 min (Collaboration et al., 9 Jul 2025).

The strategic rationale is to combine massive chemical cartography with targeted sampling of physically informative populations. DR19 explicitly states that Galactic Genesis aims to move from the APOGEE-1/2 sample of under a million unique stars to up to 3 million stars in that program alone, with disk sampling of about 100 stars pc$G-H > 3.5$2 out to 15 kpc (Collaboration et al., 9 Jul 2025). This suggests that MWM’s architecture is designed not merely for catalog growth, but for statistical reconstruction of the Galaxy’s chemodynamic history over large spatial baselines.

3. Instrumentation, pipelines, and released stellar products

MWM’s spectroscopic backbone is dual-instrument. The survey is designed to obtain high-resolution APOGEE H-band spectra and low-resolution BOSS optical spectra for stellar populations across the Milky Way (Mészáros et al., 9 Jun 2025). APOGEE provides near-infrared spectroscopy at $G-H > 3.5$3, which is particularly valuable in obscured regions of the Galactic plane, while BOSS contributes low-resolution optical spectroscopy that is often more informative for hot stars (Straumit et al., 2022).

Astra is the central analysis framework for DR19 MWM spectra. The DR19 release paper lists the principal Astra pipelines as ASPCAP, APOGEENet, AstroNN, ThePayne, BOSSNet, MdwarfType, SLAM, SnowWhite, corv, and LineForest, and it introduces astraMWMLite as a preferred-parameters summary file intended as a practical entry point into the full MWM dataset (Collaboration et al., 9 Jul 2025). For FGKM stars with $G-H > 3.5$4 K, ASPCAP remains the main APOGEE analysis engine. It pseudo-continuum normalizes the spectrum, uses FERRE to fit synthetic spectra via $G-H > 3.5$5 minimization, performs a global fit of $G-H > 3.5$6, $G-H > 3.5$7, $G-H > 3.5$8, microturbulent velocity, macroturbulent velocity for giants or $G-H > 3.5$9 for dwarfs, $1 \times 33$0, $1 \times 33$1, and $1 \times 33$2, and then fits individual elemental abundances in sensitive wavelength windows with the main parameters held fixed (Mészáros et al., 9 Jun 2025).

DR19 ASPCAP products include radial velocities, atmospheric parameters, and abundances for 24 abundance quantities corresponding to 21 elements for 964,989 stars, including 336,511 new APO observations processed for the release. The validation paper reports that the new $1 \times 33$3 values show excellent agreement with the IRFM scale, while surface gravities exhibit slight systematic offsets compared to asteroseismic gravities. The estimated precision is 50–70 K for giants and 70–100 K for dwarfs in $1 \times 33$4, 0.07–0.09 dex in $1 \times 33$5 for giants, and 0.02–0.04 dex for multiple abundances including metallicity, $1 \times 33$6, Mg, and Si. It also classifies elemental-quality regimes, with Excellent performance for $1 \times 33$7, $1 \times 33$8, C, N, O, Mg, Si, Ca, Fe, and Ni, and Poor performance for P, V, and Cu (Mészáros et al., 9 Jun 2025).

DR19 MWM quantity	Value	Context
APOGEE spectra	~1.2 million	Public MWM infrared release
BOSS spectra	~800,000	Public MWM optical release
Unique infrared stars	~390,000	APOGEE-based
Unique optical stars	~475,000	BOSS-based
ASPCAP stars	964,989	DR19 stellar-parameter and abundance products

Pipeline specialization is an important feature rather than an implementation detail. For OBAF-type stars, Zeta-Payne was built specifically for MWM and embedded in Astra as the hot-star analysis engine. It uses a neural-network emulator trained on synthetic spectra and simultaneously models the stellar spectrum and a wavelength-dependent response function. The paper concludes that for OBAF stars, BOSS optical spectra are usually more informative than APOGEE near-infrared spectra for $1 \times 33$9, $\geq 6$0, $\geq 6$1, and $\geq 6$2, although APOGEE remains useful in heavily extincted regions (Straumit et al., 2022). For cool dwarfs, a separate The Cannon 2-based M-dwarf model trained on FGK+M wide binaries produced detailed abundances for 16,590 M dwarfs with median uncertainties of 13 K in $\geq 6$3 and 0.018–0.029 dex in abundances (Behmard et al., 24 Jan 2025). MWM therefore operates less as a single pipeline than as a controlled ensemble of label-inference systems adapted to distinct spectral regimes.

4. Chemical cartography and Galactic evolution

One of the clearest demonstrations of MWM’s role in Galactic archaeology is the reconstruction of the Milky Way’s chemical map from SDSS-V DR19. In a two-infall Galactic chemical evolution analysis using OMEGA+, the survey provided a golden sample of 393,743 stars, effectively the 394,000-star abundance sample emphasized in the abstract. The analysis used $\geq 6$4, $\geq 6$5, and 14 additional elemental ratios, and treated $\geq 6$6 as the cleanest tracer of the chemically defined thin- and thick-disk bimodality (Hegedűs et al., 31 May 2025).

The abundance formalism is standard: $\geq 6$7 The chemical disk split is defined by the piecewise boundary

$\geq 6$8

Stars above this line are treated as the high-Mg / thick-disk-like sequence, and stars below as the low-Mg / thin-disk-like sequence (Hegedűs et al., 31 May 2025).

The best-fit OMEGA+ models support a two-phase history for the disk: an early rapid buildup of the thick disk / high-Mg sequence followed by a delayed second accretion episode associated with the thin disk / low-Mg sequence. For the whole disk, the best-fit timescales are a primary formation phase of $\geq 6$9 Gyr, a second-infall rising phase of $H < 12.2$0 Gyr, a relaxation after the merger of $H < 12.2$1 Gyr, and a second infall peak time of $H < 12.2$2 Gyr after Galactic birth. The same study finds a radial pattern consistent with inside-out formation, with six annuli centered at 4, 6, 8, 10, 12, and 14 kpc, a high-Mg sequence that becomes less populated in the outer disk, and a second infall time that shifts from about 4.55 Gyr at 4 kpc to about 2.67 Gyr at 14 kpc (Hegedűs et al., 31 May 2025).

A complementary line of work reframes MWM abundances in low-dimensional form. Using 70,057 red giant stars with $H < 12.2$3 and 16 elements, one study factorized the shifted abundance matrix as

$H < 12.2$4

with each star represented as a non-negative linear combination of four latent abundance patterns. The model accurately generated the measured abundances, with reduced $H < 12.2$5 for roughly 80% of stars and reduced $H < 12.2$6 for about 95%, and the four patterns were associated with early and late core-collapse supernova enrichment, Type Ia supernovae, and AGB stars (Ness et al., 19 May 2026). A second study extended this framework to 199,290 red giant stars, identifying coherent enrichment pathways that are stratified in age and height above the plane and a transition in enrichment behavior at approximately 6 Gyr (Ness et al., 20 May 2026). These results do not replace classical abundance-plane analyses; rather, they suggest that MWM’s dense multi-element chemistry supports both parametric Galactic chemical evolution modeling and data-driven latent-basis descriptions of the disk.

5. Open clusters, abundance systematics, and cluster-scale Galactic archaeology

Open clusters provide a stringent test of the internal precision and systematic control of MWM abundances. A comprehensive homogeneity study combined SDSS-V Milky Way Mapper spectra with APOGEE DR17 abundances and Gaia DR3 astrometry and kinematics to construct a clean sample of 26 open clusters with at least 6 confirmed giant members each, drawn from a parent list of approximately 2000 clusters. The analysis was restricted to giant-branch stars with $H < 12.2$7 and examined up to 20 elements/abundance dimensions, spanning $H < 12.2$8, iron-peak, odd-$H < 12.2$9, light-element, and neutron-capture families (Sinha et al., 2024).

A key result of that work is the identification of a mild but non-negligible $6 \times 67$0 trend in some APOGEE/MWM abundances. The authors corrected it cluster by cluster using

$6 \times 67$1

where $6 \times 67$2 is the fitted slope and $6 \times 67$3 sets the cluster zero point using RGB stars below the red clump. They then estimated intrinsic cluster scatter with two complementary methods: a paired-stars estimator for stars close in the HR diagram and a maximum-likelihood Gaussian intrinsic-scatter model (Sinha et al., 2024).

The headline result is that the majority of elements are homogeneous to $6 \times 67$4 dex at 99.7% confidence, while weak-line and neutron-capture elements such as Ce, Nd, Cu, V, P, S, Na are typically constrained to $6 \times 67$5 dex. Within 3$6 \times 67$6, no cluster shows measurable intrinsic inhomogeneity in any element. The most precise pair-based measurements reach scatter limits at the $6 \times 67$7 dex level, and giant stars in open clusters are on average $6 \times 67$8 dex more chemically homogeneous than matched field stars, although the difference is generally consistent with zero within $6 \times 67$9 (Sinha et al., 2024). The paper interprets these results as support for very efficient mixing in progenitor molecular clouds and limited ISM pollution during cluster formation.

At larger scale, the OCCAM DR19 analysis used MWM/DR19 spectroscopy plus Gaia astrometry and kinematics to build a sample of 164 high-quality open clusters and 1083 member stars, a 74% increase in high-quality clusters relative to the previous OCCAM DR17 release. Using these clusters as calibrated tracers of Galactic chemical evolution, the study measured a radial metallicity gradient

$12 \times 67$0

and, using guiding-center radius,

$12 \times 67$1

It also reported mostly shallow or statistically insignificant $12 \times 67$2 gradients for many elements and found only tentative evidence for azimuthal variation in the radial metallicity gradient (Otto et al., 9 Jul 2025). Together, the homogeneity and OCCAM results show that MWM is capable of using clusters both as precision calibration laboratories and as spatial tracers of Galactic abundance structure.

6. Extended stellar populations, ISM cartography, and survey boundaries

Although much MWM Galactic archaeology is built on red giants, the survey’s scientific reach is broader. The M-dwarf abundance study demonstrates that APOGEE H-band spectra can be repurposed for detailed chemistry in cool dwarfs at scale, using a data-driven model trained on 79 FGK+M binaries and applied to 16,590 M dwarfs in SDSS-V/MWM (Behmard et al., 24 Jan 2025). The hot-star pipeline study shows that MWM also required dedicated machinery for OBAF stars, because no established survey pipelines existed for a substantial fraction of the spectra, estimated at about 10% of the MWM total (Straumit et al., 2022). In both cases, the technical message is that MWM is a heterogeneous stellar-survey platform rather than a single-population abundance program.

MWM-like spectroscopy is also relevant beyond stellar labels. A three-dimensional dust-mapping study based on APOGEE-2 near-infrared spectroscopy used more than 44,000 stars with Bayesian and Gaussian-process reconstruction to produce a Galactic-plane dust map extending 10 kpc from the Sun, with effective spatial resolution of about 100 pc, and a catalog of 84 large molecular clouds with non-kinematic distances and volume densities (Kh. et al., 2024). The paper presents this as a demonstration of how APOGEE-quality near-infrared stellar spectroscopy can be turned into a three-dimensional ISM tomography tool. This suggests a broader interpretation of MWM: not only as a stellar-survey program, but also as an enabling infrastructure for Galactic cartography in dust and star-forming structure.

A final boundary condition concerns comparison with DESI’s stellar survey. The DESI Milky Way Survey (MWS) is a distinct program that will observe approximately seven million stars over $12 \times 67$3 in bright time, with a deliberately inclusive and forward-modelable low-resolution optical selection. Its main sample is divided into main-blue, main-red, and main-broad, and it also includes high-priority classes such as white dwarfs, nearby stars within 100 pc, RR Lyrae, and BHB stars (Cooper et al., 2022). Earlier target-selection documentation described the scope as $12 \times 67$4 million stars between $12 \times 67$5 mag, supplemented by brighter targets in poor observing conditions (Prieto et al., 2020). The scientific themes overlap strongly with those of MWM, but the surveys differ in formal identity, resolution, wavelength coverage, and target-selection architecture. Recognizing that distinction avoids conflating SDSS-V MWM abundance products with DESI MWS radial-velocity and low-resolution metallicity products.

In aggregate, the published literature portrays MWM as a survey system rather than a single catalog. Its defining characteristics are survey-scale optical and infrared spectroscopy, explicit targeting and selection-function infrastructure, multiple specialized inference pipelines within Astra, and a scientific program that ranges from stellar-parameter validation and cluster homogeneity to disk-wide chemical evolution, latent-pattern decomposition, cool-dwarf abundance inference, hot-star spectroscopy, and dust tomography (Collaboration et al., 9 Jul 2025, Mészáros et al., 9 Jun 2025).