Into the Darkness: Classical and Type II Cepheids in the Zona Galactica Incognita

Eva K. Grebel; Gergely Hajdu; Istv\'an D\'ek\'any; M\'arcio Catelan

REVIEW 3 major objections 5 minor 131 references

Into the Darkness: Classical and Type II Cepheids in the Zona Galactica Incognita

T0 review · 3 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read This paper reports the discovery of 640 distant classical Cepheids and more than 500 type II Cepheids in the heavily obscured inner Galaxy and far side of the Milky Way disk, and uses them to measure the near-infrared extinction law and…

desk verdict A major far-side Cepheid catalog, but its distance scale rests on a single extinction-calibration assumption that deserves a stress test. read the letter →

arxiv 1908.08290 v1 pith:OVKYWC27 submitted 2019-08-22 astro-ph.SR astro-ph.GAastro-ph.IM

István Dékány , Gergely Hajdu , Eva K. Grebel , Márcio Catelan This is my paper · ORCID

classification astro-ph.SRastro-ph.GAastro-ph.IM

keywords DeltaCepheidvariablestarsTypeIICepheidsCatalogsSurveysGalacticbulgeMilkyWaydiskInterstellarextinctionconvolutionalneuralnetwork

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper aims to open the "Zona Galactica Incognita," the heavily obscured far side of the Milky Way's disk, to stellar mapping. Using near-infrared time-series photometry of the southern Galactic plane, it discovers 640 new distant classical Cepheids, some behind up to roughly 40 magnitudes of visual extinction, and more than 500 type II Cepheids, most in the inner bulge. A convolutional neural network classifies the light curves, and a second neural network predicts the pulsation-phase color corrections that turn sparse photometry into unbiased reddening estimates. From these it derives a steep, spatially varying near-infrared extinction curve toward the bulge and calibrates the selective-to-absolute extinction ratios by aligning the bulge type II Cepheids' distance peak with the known Galactic-center distance. If the calibration holds, the far side of the disk now has stellar distance indicators, and the claimed warp, flaring, and radial age gradient of the thin disk follow.

What carries the argument

The central machinery is a convolutional neural network that treats a phase-folded, phase-aligned, standardized $K_s$-band light curve as a one-dimensional image, with the pulsation period and amplitude added as extra input channels; it classifies candidates as classical Cepheid, type II Cepheid, or non-Cepheid. A second neural network predicts the phase-dependent color corrections $\Delta(J-K_s)$ and $\Delta(H-K_s)$ from the $K_s$-band light-curve parameters, so that sparse $J$ and $H$ photometry can be converted into unbiased mean colors. What carries the distance argument is the near-infrared period-luminosity relation, the tight empirical connection between a Cepheid's pulsation period and its absolute brightness: it fixes the absolute magnitude from the period, so a measured $K_s$ magnitude and color give the extinction and the distance once the selective-to-absolute extinction ratio is known. That ratio is pinned down by assuming the bulge type II Cepheids are centrally symmetric around the Galactic center and tuning the ratio until the peak of their line-of-sight distance distribution matches the known Galactic-center distance.

What would settle it

Measure distances for the bright bulge type II Cepheids with independent geometric or trigonometric parallaxes and compare the peak of their line-of-sight distance distribution with the established Galactic-center distance of 8178 pc; a mismatch larger than the quoted uncertainties would show that the symmetry assumption used to set the extinction ratios is wrong.

Watch

Extended reading notes

Core claim

The central discovery is that stellar distance indicators can be found and used in the most obscured part of the Galaxy, the far side of the disk and the inner bulge. From five years of near-infrared time-series photometry of the southern Galactic plane, the paper reports 689 classical Cepheids (640 of them new) with extinctions reaching about 40 magnitudes in the visual, and 608 type II Cepheids, most in the bulge. A convolutional neural network separates the two classes from $K_s$-band light curves, and the authors estimate about 10% contamination in each sample. Using neural-network predictions of pulsation-phase color variation, the Cepheids become reddening tracers; the bulge type II Cepheids then yield mean extinction ratios $A(K_s)/E(J-K_s)=0.528\pm0.004$ (stat.) $\pm0.019$ (sys.) and $A(K_s)/E(H-K_s)=1.50\pm0.01$ (stat.) $\pm0.05$ (sys.), and reveal a near-infrared extinction curve that varies on roughly $5^\circ$ scales. With those distances, the classical Cepheids trace a warped, flared outer disk and a radial age gradient, while the type II Cepheids trace a centrally concentrated, slightly elongated old bulge population.

Load-bearing premise

The calibration of how much total extinction corresponds to a measured reddening assumes that the bulge type II Cepheids are spread symmetrically around the Galactic center along our line of sight; if their true distribution is lopsided or tilted with a density gradient, those extinction ratios and every distance built on them shift.

Editorial extensions

If this is right

The far side of the Galactic disk now has stellar distance indicators out to roughly 20 kpc, so maps of the warp and flare no longer rely only on gas kinematics or near-side tracers.
The mean near-infrared reddening ratio $E(J-K_s)/E(H-K_s)\simeq2.83$ agrees between the bulge and disk footprints, yet varies by about 2% on angular scales of about $5^\circ$ toward the bulge; distance work in the inner Galaxy therefore needs a spatially resolved extinction law.
Bulge type II Cepheids form a centrally concentrated, slightly elongated old population whose inclination matches the inner RR Lyrae distribution, supporting a radius-dependent orientation of the old bulge.
Cepheid ages in the far disk show a radial and vertical gradient: stars younger than about 70 Myr concentrate inside the Solar circle, while stars older than about 120 Myr are found outside it and farther from the plane.
With the new extinction law, only 9 classical Cepheids lie within 3 kpc of the Galactic center and 3 within 2 kpc, so the earlier claim of a young disk crossing the inner Milky Way is no longer clearly supported.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

This suggests that applying the same convolutional-network classification and neural color-correction to near-infrared time-domain surveys of the northern mid-plane would turn the warp and flare measurements into a full 360-degree map of the disk, testable against gas-based spiral models.
If the spatial variation of the near-infrared extinction law on $5^\circ$ scales is real, single-band distance estimates in the inner Galaxy carry an irreducible systematic error; this could be checked by comparing Cepheid distances with future precise parallaxes for a subsample.
The symmetry-based extinction calibration could be stress-tested with a simulated triaxial bar: a tilted bar with a density gradient along the line of sight would bias the fitted extinction ratios, and the size of that bias could be quantified without new observations.
The neural color-correction method should transfer to RR Lyrae stars and Miras, turning other pulsators into unbiased reddening tracers in crowded, highly extincted fields.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Two things to know before you read it. First, this is a real observational advance: 640 previously unknown classical Cepheids and over 500 type II Cepheids in the far-side disk and bulge, found with a genuinely new CNN classifier for near-IR light curves and a neural-net color-correction scheme. If the distances hold, it roughly triples the Cepheid census in the most obscured parts of the Galaxy and gives the first stellar tracer view of the warp, flare, and age gradients on the far side. Second, the entire distance scale rests on one load-bearing assumption that gets less scrutiny than it deserves: the Sect. 4.3 extinction calibration.

The method tunes the KDE peak of the bulge type II Cepheid dY distribution to GRAVITY's R0 = 8178 pc, assuming a centrally symmetric intrinsic distribution. That assumption is insecure for two reasons. The bulge/bar is triaxial and tilted, so a density gradient along the sightline shifts the mode. And the survey's own motivation is that extinction rises toward and beyond the Galactic center, so far-side Cepheids are fainter and less complete; the observed mode is pulled to the near side. Both effects force RKJK upward when you tune to R0, which inflates the extinction corrections and pushes the classical Cepheid distances outward—exactly the stars driving the far-side structure claims. The quoted ±0.019 systematic comes from Monte Carlo resampling of the same estimator, so it cannot capture a biased mode. The earlier Dékány/Matsunaga disagreement is a concrete reminder that a ~10% ratio difference can flip conclusions about the inner Galaxy.

The rest of the paper is more solid. The photometric recalibration is careful, the CNN is tested with synthetic noise-injected light curves and checked against OGLE classifications, and the spatial variation analysis of the reddening ratio is honest—the authors explicitly warn against interpolating their binned results. The training set does include 48 of 188 classical Cepheids selected using the authors' own distance-extinction consistency method, which introduces some circularity, but they argue the selection is insensitive to the extinction law and the objects were visually inspected. Not fatal, but worth noting. The quoted precision on the reddening ratio (2.832 ± 0.004) also looks optimistic given the scatter, and the paper ships no code, which makes the ML parts harder to reproduce.

Who's it for: anyone working on Galactic structure, extinction, or Cepheid populations. It deserves a serious referee and, I think, publication after the authors stress-test the Sect. 4.3 calibration—for example with a completeness model, or by comparing against an independent distance anchor like RR Lyrae stars. The catalog and the extinction-variation maps are valuable regardless of the exact distance scale.

Referee Report

3 major / 5 minor

Summary. This paper presents a near-infrared census of classical and type II Cepheids in the VVV southern disk and inner-bulge footprint. The authors recalibrate the VVV photometry, build a convolutional-neural-network classifier on phase-folded Ks-band light curves using a training set of 188 classical and 356 type II Cepheids, apply it to about 40,000 candidates, visually inspect the outputs, and obtain a final sample of 689 classical Cepheids (640 new) and 608 type II Cepheids (over 500 in the inner bulge). They predict pulsation-phase color corrections with neural networks, derive the reddening ratio R_JKHK = 2.832 ± 0.004 and the selective-to-absolute extinction ratios R_KJK = 0.528 and R_KHK = 1.50 (Eqs. 22 and 24), compute heliocentric distances, and use the resulting three-dimensional distribution to trace the bulge population, the Galactic warp and flare, and radial and vertical age gradients in the far-side disk.

Significance. If the distances are reliable, this is a landmark data set: it populates the previously obscured far side of the Milky Way disk with standard candles, provides new constraints on the near-infrared extinction curve in the inner Galaxy, and offers the first Cepheid-based view of the warp, flare, and age structure beyond the Galactic center. The paper is unusually strong in data products and method transparency, with machine-readable light curves and catalogs, a careful discussion of VVV photometric zero-point problems, and a synthetic-noise validation of the classifier. The central scientific payoff, however, rests on the Sect. 4.3 extinction-ratio calibration, whose symmetry and completeness assumptions are not validated against plausible lopsided bulge models or the survey's own strong extinction gradient. The quoted systematic uncertainties are internal to the estimator and cannot capture a biased distance mode, so the distance scale and all downstream spatial-structure conclusions require a robustness test before the census can be taken at face value.

major comments (3)

[Sect. 4.3, Eqs. (22)-(24), Table 8] The calibration of R_KJK = A(Ks)/E(J-Ks) is obtained by forcing the KDE peak of dY = d_H cos b cos l for bulge type II Cepheids to R0 = 8178 pc. This is valid only if the underlying three-dimensional distribution is centrally symmetric about the Galactic center and if the observed sample is unbiased in distance. Both assumptions are questionable in the inner Galaxy: a lopsided or tilted bar/density asymmetry shifts the distance mode, and the survey's own extinction map implies that far-side bulge Cepheids are preferentially missed, pulling the observed dY mode to the near side. Tuning the observed mode to R0 then changes the inferred R_KJK; for the completeness effect alone, the fit would likely bias R_KJK low rather than high, because a smaller A(Ks) is needed to push near-side stars outward. Equations (22) and (24) enter every distance in Table 8 and therefore all warp, flare, and age-gradient results in Sect. 5. The ±0.019 systematic in Eq. (22) comes from Monte Carlo resampling of the same estimator and cannot capture this mode bias. I request a forward-model test: generate mock bulge type II Cepheids from a triaxial or lopsided density model, apply the magnitude- and position-dependent VVV completeness function, run the same KDE/R0 fitting procedure, and report the resulting bias in R_KJK. If this bias is comparable to or larger than 0.019, the systematic errors and the distances derived from them need to be enlarged accordingly.
[Sect. 3.3 and Sect. 3.8 (training set and final sample)] Forty-eight of the 188 classical Cepheid training examples were selected by the authors' tentative distance-extinction consistency method (delta >= 3, d < 7.5 kpc, Sect. 3.3 and Fig. 4), using the same PL relations, extinction-map assumptions, and reddening framework that are later used to derive distances in Sects. 4.2-4.3. Because the final DCEP/T2CEP classification of the survey data is made by a CNN trained on this set, any systematic error in that selection can imprint itself on the census of 640 new classical Cepheids in a way that cross-validation accuracy cannot detect, since the same assumptions are embedded in the labels. The authors state that the selection is insensitive to the extinction law within the range considered, but a direct test would be more convincing: retrain the CNN without the 48 internally selected objects and compare the classifications and final counts on the full candidate sample. If the census changes materially, the overlap between the training selection and the distance analysis should be disclosed and discussed as a systematic limitation.
[Sect. 5.2, Fig. 22] The claimed significant vertical and radial age gradients rest on period-age relations from Anderson et al. (2016), evaluated at metallicities assigned from a near-side radial metallicity gradient, with no individual metallicities or rotation/instability-strip information for the far-side Cepheids. The authors acknowledge the P ~ 10 d classification confusion and the modeling assumptions, but the age gradient is presented as one of the main discoveries. I ask for a robustness test: recompute the median-age curves with a +/-0.2 dex shift in the adopted [Fe/H] and with the alternative instability-strip-crossing and rotation choices in Anderson et al. (2016), and state whether the gradient slope and significance survive.

minor comments (5)

[Sect. 3.2 and Sect. 3.7] There are typographical errors: 'stranderdized' in Sect. 3.2 should be 'standardized', and 'in oder' in Sect. 3.7 should be 'in order'.
[Sect. 4.2, Fig. 14] The text says the quoted errors on R_JKHK include all statistical and systematic uncertainties, but the binned spatial-variation analysis in the same section explicitly uses statistical errors only; please clarify which components are included in the quoted +/-0.004 and in the binned values shown in Fig. 15.
[Sect. 3.5, Fig. 8] The synthetic-noise test draws noise realizations added to signals from the training set itself, so the performance curves may be slightly optimistic; this is stated in the text, but it would be useful to add the same caveat directly in the figure caption.
[Sect. 3.7] The OGLE comparison is based on only 41 common objects and the estimated recall of 0.92 should be quoted with a binomial confidence interval rather than as a point value, given the small sample size.
[Table 8 and Fig. 19] Objects with dH > 40-50 kpc and small A(Ks) are explicitly suspected to be misclassified; consider adding a classification-quality or distance-quality flag to the electronic catalog so that downstream users do not treat all 689 distances as being of uniform reliability.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the extinction scale is anchored to the external GRAVITY R0 and independent PL relations, and the central-symmetry assumption is explicit rather than concealed.

full rationale

The paper's derivation chain is: CNN classification (with literature/OGLE training plus 48 authors' candidates) -> mean colors via neural color-correction -> color excesses from PL relations -> reddening ratio R_JKHK (a slope of E(J-Ks) versus E(H-Ks), independent of the absolute extinction scale) -> selective-to-absolute ratios by forcing the bulge type II Cepheid dY peak to the GRAVITY-measured R0 = 8178 pc -> distances via Eq. 25. The calibration step in Sect. 4.3 is an explicit calibration to an external anchor, not a prediction masquerading as a test. The central-symmetry assumption is stated openly and is not derived from the data; the resulting shape and elongation of the T2C distribution retain independent content. The 48 classical Cepheids added to the CNN training set were selected by the authors' earlier distance-extinction consistency method, and the paper explicitly checks that the same selection is obtained with the later extinction ratios and validates the classifier against OGLE classifications; this is a selection detail, not a reduction of the later extinction or distance measurements to the training labels. Self-citations (Dekany et al. 2013, 2015b, 2018; Hajdu et al. 2019) are methodological or contextual and are not load-bearing circular justifications. No equation in the paper is equivalent to its input by construction, and no fitted parameter is renamed as a prediction. The main scientific claims (640 new classical Cepheids, warp/flare, age gradients) depend on the calibrated extinction law but are not identical to the calibration input. Therefore no circular step is identified.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The central distance-dependent results rest on a small number of externally calibrated relations and one geometric assumption. The near-IR extinction coefficients are fitted quantities, not first-principles derivations. The main internal consistency risk is the use of the same type II Cepheid sample for both extinction calibration and bulge/distance science, and the use of self-selected classical Cepheid training examples.

free parameters (3)

RKJK = A(Ks)/E(J-Ks) = 0.528 +/- 0.004 (stat) +/- 0.019 (sys)
Fitted in Sect. 4.3 by varying the value until the kernel density peak of the type II Cepheid line-of-sight distance distribution matches the GRAVITY R0. Used in Eq. 22 and to compute all distances.
RJKHK = E(J-Ks)/E(H-Ks) = 2.832 +/- 0.004 (bulge); 2.833 +/- 0.004 (disk)
Fitted by Bayesian linear regression to the color excess distribution in Sect. 4.2. Combined with RKJK to derive RKHK.
RKHK = A(Ks)/E(H-Ks) = 1.50 +/- 0.01 (stat) +/- 0.05 (sys)
Adopted in Sect. 4.3 (Eq. 24) by combining the mean RJKHK with the fitted RKJK; used with RKJK to compute distances.

assumptions (5)

domain assumption Type II Cepheids in the bulge are centrally symmetric about the Galactic center (Sgr A*).
Used in Sect. 4.3 to calibrate the extinction coefficients by requiring the distance distribution to peak at R0 = 8178 pc. If the distribution is lopsided along the sightline, the coefficients and all derived distances shift.
domain assumption The adopted near-infrared period-luminosity relations (Macri et al. 2015; Bhardwaj et al. 2017, transformed to VISTA) are valid for the surveyed Cepheids.
Equations 13-21. The zero-points are adjusted to the LMC distance modulus; errors in these relations propagate into all distances and extinction estimates.
domain assumption The period-age relations of Anderson et al. (2016), with metallicities from the Genovali et al. (2014) and Luck (2018) radial gradients, apply to these Cepheids.
Used in Sect. 5.2 to compute individual ages. Rotation and instability-strip crossing are averaged over, and individual metallicities are unknown.
domain assumption The recalibrated VVV photometry (Hajdu et al. 2019) is free of systematic zero-point errors at the level claimed.
The entire analysis depends on the recalibrated photometric zero-points described in Sect. 2.2; residual zero-point errors would bias colors and extinctions.
domain assumption The CNN training set, including 48 self-selected VVV classical Cepheids, is representative of the survey's target population.
Sect. 3.3. The data-mismatch test uses synthetic noise added to the same labeled examples, so the ~10% contamination estimate may be optimistic.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Into the Darkness: Classical and Type II Cepheids in the Zona Galactica Incognita." pith.science (2026). https://pith.science/paper/OVKYWC27

@misc{pith2026190808290,
  author       = {Pith},
  title        = {Pith review of: Into the Darkness: Classical and Type II Cepheids in the Zona Galactica Incognita},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/OVKYWC27}},
  note         = {Machine review of arXiv:1908.08290}
}

read the original abstract

The far side of the Milky Way's disk is one of the most concealed parts of the known Universe due to extremely high interstellar extinction and point source density toward low Galactic latitudes. Large time-domain photometric surveys operating in the near-infrared hold great potential for the exploration of these vast uncharted areas of our Galaxy. We conducted a census of distant classical and type II Cepheids along the southern Galactic mid-plane using near-infrared photometry from the VISTA Variables in the V\'ia L\'actea survey. We performed a machine-learned classification of the Cepheids based on their infrared light curves using a convolutional neural network. We have discovered 640 distant classical Cepheids with up to ~40 magnitudes of visual extinction, and over 500 type II Cepheids, most of them located in the inner bulge. Intrinsic color indices of individual Cepheids were predicted from sparse photometric data using a neural network, allowing their use as accurate reddening tracers. They revealed a steep, spatially varying near-infrared extinction curve toward the inner bulge. Type II Cepheids in the Galactic bulge were also employed to measure robust mean selective-to-absolute extinction ratios. They trace a centrally concentrated spatial distribution of the old bulge population with a slight elongation, consistent with earlier results from RR Lyrae stars. Likewise, the classical Cepheids were utilized to trace the Galactic warp and various substructures of the Galactic disk, and to uncover significant vertical and radial age gradients of the thin disk population at the far side of the Milky Way.

Figures

Figures reproduced from arXiv: 1908.08290 by the authors.

**Figure 1.** Ks-band light curves of the same Cepheid toward the central bulge, phase-folded with the best-fitting period shown in the upper left corner. The left panel shows photometry calibrated by CASU, the right panel displays recalibrated photometry according to Hajdu et al. (2019). increasing period. This bump is a characteristic light curve detail that can be used to distinguish some classical Cepheids from other types o… view at source ↗

**Figure 2.** Fourier parameters of the I-band light curves of fundamental-mode classical Cepheids in the Large Magellanic Cloud, observed by the OGLE survey (Soszy´nski et al. 2015). (e.g., Krizhevsky et al. 2012; Simonyan, & Zisserman 2014). At the same time, CNNs have been also successfully employed in one-dimensional (sequence) data in diverse fields ranging from natural language processing (e.g., Dauphin et al. 2016) to m… view at source ↗

**Figure 3.** Phase-folded, phase-aligned and standardized light curve models of all classical Cepheids in our training set (gray curves). Black points denote the light curve model evaluated in 38 equidistant phase points for one training example, which serves as an input sequence of our CNN classifier (see Sect. 3.1). that their extinctions and distances become consistent under the classical Cepheid assumption). We selected th… view at source ↗

Figures from the paper (21 more)

**Figure 4.** Figure 4: Weighted standard deviations of the absolute Ks-band extinction computed from the VVV photometry (see text) with respect to the corresponding extinction values from Gonzalez et al. (2012) as a function of Heliocentric distance for our bulge Cepheid candidates (black p…

**Figure 5.** Figure 5: Summary of the CNN architectures evaluated in our model selection procedure, with data propagating from bottom to top (see text for details). Each rectangle represents a building block of the model, and their tunable hyperparameters are listed in italics. Optional mo…

**Figure 6.** Figure 6: Left: training (TR, red) and cross-validation (CV, green) loss as a function of training epochs. Semi-transparent thin curves show the loss per fold, thick curves show their mean. Right: classification accuracy of the training and validation sets as a function of train…

**Figure 7.** Figure 7: Confusion matrix of our CNN classifier measured by 10-fold cross-validation [PITH_FULL_IMAGE:figures/full_fig_p012_7.png]

**Figure 8.** Figure 8: Left panel: performance estimates as a function of mean apparent Ks magnitude, using a bin size of 0.5 mag. Right panel: as in left panel but showing cumulative distributions. The figure keys have the following notation: A: overall accuracy; P, R, F1, A’: binary precis…

**Figure 9.** Figure 9: Light curve of an eclipsing binary star misclassified as a classical Cepheid by our CNN classifier. The photometric measurements are phase-folded with the (incorrect) period found by our algorithm (left panel) and its doubled value (right panel). The red curve shows th…

**Figure 10.** Figure 10: Distribution of objects classified as type II Cepheids (top) and classical Cepheids (bottom) in our final sample, shown in the Galactic coordinate system. Individual objects are represented by gray dots, the color scale represents their kernel density estimate [PITH_…

**Figure 11.** Figure 11: Histogram of the periods of the objects classified as classical Cepheids (DCEP, filled bars) or type II Cepheids (T2CEP, step curves) shown on a logarithmic scale. The blue and the red colors represent subsamples from the VVV survey’s bulge and disk footprints, respe…

**Figure 12.** Figure 12: Gray points: normalized and phase-aligned color curves of the classical Cepheids (left and middle) and type II Cepheids (right) in the training sets of our predictive models of the color index. Black points: mean absolute cross-validation error of our predictive model…

**Figure 13.** Figure 13: Phase-folded Ks-band light curves (upper panels) and J − Ks color index curves (lower panels) of two classical Cepheids from our disk sample. In the upper panels, black symbols mark the observational data and their errors, red curves denote the fitted light curve mode…

**Figure 15.** Figure 15: Spatial variations in the RJKHK reddening ratio as a function of Galactic longitude (top) and latitude (bottom) over the bulge area of our study, using bin sizes of 5 ◦ and 0.5 ◦ , respectively. inating from the PL relations would affect all binned subsamples in the s…

**Figure 16.** Figure 16: Histogram of the dY distance components (see text.) of the bulge type II Cepheid sample with log P < 1.3 and the corresponding kernel density estimate (black curve, arbitrarily scaled up to match the histogram) for a trial value of RKJK. 0.46 0.48 0.50 0.52 0.54 0.56 …

**Figure 17.** Figure 17: Variation of the Rˆ0 as a function of the RKJK (see text) are shown by black curves, each curve representing one realization of the Monte Carlo simulation of statistical errors. The green shaded area shows the total uncertainty around the value of R0 by Gravity Col…

**Figure 18.** Figure 18: Kernel density estimate of the projected distance distribution of type II Cepheids in the Galactic bulge from latitudinal ranges shown in the figure, represented as a color scale and contour lines. The white dashed lines show the l = ±2 ◦ and l = ±5 ◦ sight-lines. Th…

**Figure 19.** Figure 19: Histogram of the Heliocentric distances of classical Cepheids on a logarithmic scale. of dH > 40 kpc, and only 8% of them have dH > 50 kpc. Most of these objects with anomalously large distances have small S/N (. 100), and are most probably misclassified. The latter …

**Figure 20.** Figure 20: Top panels: histograms of the Ks-band extinction (left) and the apparent mean Ks magnitude of the classical Cepheids in our sample. Bottom panels: distributions of the Z distance from the Galactic plane marginalized over the projected Heliocentric (dH,proj., left pane…

**Figure 21.** Figure 21: Left: the spatial distributions of the classical Cepheids in our sample projected onto the Galactic plane. The red ‘x’ symbol marks the position of the Galactic center. Crosses and diamonds of different sizes indicate how far above or below, respectively, the objects …

**Figure 22.** Figure 22: Left panel: black points denote median ages of classical Cepheids in the Galactic disk computed for 10 subsamples in 2 kpc-wide overlapping bins of Galactocentric cylindrical distance. The vertical bars show error estimates of the median, the horizontal bars mark the …

**Figure 23.** Figure 23: Spatial distribution of the classical Cepheids projected onto the Galactic plane. Black and red points show the positions of Cepheids with P ≤ 10 d and P > 10 d, respectively. The color scale and the white contours denote a kernel density estimate of the distribution …

**Figure 24.** Figure 24: Comparison of the mean apparent Ks magnitudes (left), H − Ks color indices (middle), and Ks-band absolute extinctions (right) of the D´ek´any et al. (2015b) Cepheid sample as presented in that study (denoted as ‘old’) vs our present analysis (denoted as ‘new’). 0.45 0…

**Figure 25.** Figure 25: The mean selective-to-absolute extinction ratio derived from the Cepheid sample in this paper, in comparison with literature values from Nishiyama et al. (2009, N09), Indebetouw et al. (2005, I05), Chen et al. (2018, C18) and Schlafly et al. (2016, S16, in conjuncti…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

131 extracted references · 65 canonical work pages

[1]

, " * write output.state after.block = add.period write newline

ENTRY address archivePrefix author booktitle chapter doi edition editor eprint howpublished institution journal key month number organization pages publisher school series title misctitle type volume year version url label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts ...
[2]

write newline

" write newline "" before.all 'output.state := FUNCTION format.url url empty "" new.block "" url * "" * if FUNCTION format.eprint eprint empty "" archivePrefix empty "" archivePrefix "arXiv" = new.block " " eprint * " " * new.block " " eprint * " " * if if if FUNCTION format.doi doi empty "" " " doi * " " * if FUNCTION format.pid doi empty eprint empty ""...
[4]

Abadi, M., Agarwal, A., Barham, P., et al.\ 2016, arXiv e-prints, arXiv:1603.04467

arXiv 2016
[5]

Gravity Collaboration, Abuter, R., Amorim, A., et al.\ 2019, , 625, L10

2019
[6]

Abuter, R., Amorim, A., Bauboeck, M., et al.\ 2019, arXiv e-prints , arXiv:1904.05721

arXiv 2019
[7]

Anders, F., Khalatyan, A., Chiappini, C., et al.\ 2019, arXiv e-prints, arXiv:1904.11302

arXiv 2019
[8]

I., Eyer, L., & Mowlavi, N.\ 2013, , 434, 2238

Anderson, R. I., Eyer, L., & Mowlavi, N.\ 2013, , 434, 2238

2013
[9]

I., Saio, H., Ekstr \"o m, S., et al.\ 2016, , 591, A8

Anderson, R. I., Saio, H., Ekstr \"o m, S., et al.\ 2016, , 591, A8

2016

Show all 131 references

[10]

Alonso-Garc \' a, J., Minniti, D., Catelan, M., et al.\ 2017, , 849, L13

2017
[11]

G., III, Fernley, J

Barnes, T. G., III, Fernley, J. A., Frueh, M. L., et al.\ 1997, , 109, 645

1997
[12]

M., Rejkuba, M., et al.\ 2017, , 153, 154

Bhardwaj, A., Macri, L. M., Rejkuba, M., et al.\ 2017, , 153, 154

2017
[13]

Bhardwaj, A., Rejkuba, M., Minniti, D., et al.\ 2017, , 605, A100

2017
[14]

Bono, G., Marconi, M., Cassisi, S., et al.\ 2005, , 621, 966

2005
[15]

F., Bhardwaj, A., Contreras Ramos, R., et al.\ 2018, , 619, A51

Braga, V. F., Bhardwaj, A., Contreras Ramos, R., et al.\ 2018, , 619, A51

2018
[16]

A., Coleman, T

Branch, M. A., Coleman, T. F., Li, Y.\ 1999, SIAM Journal on Scientific Computing, 21, 1

1999
[17]

Brunel, A., Pasquet, J., Pasquet, J., et al.\ 2019, arXiv e-prints , arXiv:1901.00461

2019 arXiv
[18]

Camargo, D., Bica, E., & Bonatto, C.\ 2013, , 432, 3349

2013
[19]

A.\ 2015, Pulsating Stars (Wiley-VCH)

Catelan, M., & Smith, H. A.\ 2015, Pulsating Stars (Wiley-VCH)

2015
[20]

Chen, X., Wang, S., Deng, L., & de Grijs, R.\ 2018, , 859, 137

2018
[21]

Chen, X., Wang, S., Deng, L., et al.\ 2018, , 237, 28

2018
[22]

Chen, X., Wang, S., Deng, L., et al.\ 2019, Nature Astronomy, 3, 320

2019
[23]

Chen, B.-Q., Huang, Y., Hou, L.-G., et al.\ 2019, , 487, 1400

2019
[24]

https://github.com/fchollet/keras

Chollet, F.\ 2015, keras, GitHub. https://github.com/fchollet/keras

2015
[25]

Clevert, D.-A., Unterthiner, T., & Hochreiter, S.\ 2015, arXiv e-prints , arXiv:1511.07289

2015 arXiv
[26]

N., Fan, A., Auli, M., et al.\ 2016, arXiv e-prints , arXiv:1612.08083

Dauphin, Y. N., Fan, A., Auli, M., et al.\ 2016, arXiv e-prints , arXiv:1612.08083

2016 arXiv
[27]

M., Marshall, J

Davies, T. M., Marshall, J. C., Hazelton, M. L.\ 2018, Statistics in Medicine, 37, 1191

2018
[28]

M., Aerts, C., et al.\ 2007, , 475, 1159

Debosscher, J., Sarro, L. M., Aerts, C., et al.\ 2007, , 475, 1159

2007
[29]

D \'e k \'a ny, I., Minniti, D., Catelan, M., et al.\ 2013, , 776, L19

2013
[30]

D \'e k \'a ny, I., Minniti, D., Hajdu, G., et al.\ 2015, , 799, L11

2015
[31]

D \'e k \'a ny, I., Minniti, D., Majaess, D., et al.\ 2015, , 812, L29

2015
[32]

K., et al.\ 2018, , 857, 54

D \'e k \'a ny, I., Hajdu, G., Grebel, E. K., et al.\ 2018, , 857, 54

2018
[33]

Dobbs, C., & Baba, J.\ 2014, , 31, e035

2014
[34]

R., Wang, Q

Dong, H., Mauerhan, J., Morris, M. R., Wang, Q. D., & Cotera, A.\ 2015, , 446, 842

2015
[35]

and Hart, P.\ 1973, Pattern Classification and Scene Analysis, 271 (John Wiley and Sons)

Duda, R. and Hart, P.\ 1973, Pattern Classification and Scene Analysis, 271 (John Wiley and Sons)

1973
[36]

Elorrieta, F., Eyheramendy, S., Jord \'a n, A., et al.\ 2016, , 595, A82

2016
[37]

P., Irwin, M

Emerson, J. P., Irwin, M. J., Lewis, J., et al.\ 2004, , 5493, 401

2004
[38]

W., Menzies, J

Feast, M. W., Menzies, J. W., Matsunaga, N., & Whitelock, P. A.\ 2014, , 509, 342

2014
[39]

Feldmeier-Krause, A., Neumayer, N., Sch \"o del, R., et al.\ 2015, , 584, A2

2015
[40]

L., & Massa, D.\ 2009, , 699, 1209

Fitzpatrick, E. L., & Massa, D.\ 2009, , 699, 1209

2009
[41]

B., Stead, J

Foster, J. B., Stead, J. J., Benjamin, R. A., Hoare, M. G., & Jackson, J. M.\ 2012, , 751, 157

2012
[42]

Genovali, K., Lemasle, B., Bono, G., et al.\ 2014, , 566, A37

2014
[43]

Glorot, X., Bordes, A., Bengio, Y.\ 2011, Proceedings of Machine Learning Research 15, 315

2011
[44]

A., Rejkuba, M., Zoccali, M., et al.\ 2012, , 543, A13

Gonzalez, O. A., Rejkuba, M., Zoccali, M., et al.\ 2012, , 543, A13

2012
[45]

T., Irwin, M

Gonz \'a lez-Fern \'a ndez, C., Hodgkin, S. T., Irwin, M. J., et al.\ 2018, , 474, 5459

2018
[46]

J.\ 1998, , 85, 161

Grevesse, N., & Sauval, A. J.\ 1998, , 85, 161

1998
[47]

K., & Jurcsik, J.\ 2018, , 857, 55

Hajdu, G., D \'e k \'a ny, I., Catelan, M., Grebel, E. K., & Jurcsik, J.\ 2018, , 857, 55

2018
[48]

K.\ 2019, arXiv:1908.06160

Hajdu, G., D \'e k \'a ny, I., Catelan, M., & Grebel, E. K.\ 2019, arXiv:1908.06160

2019 arXiv
[49]

J., Tibshirani, R

Hastie, T. J., Tibshirani, R. J., Friedman, J. H.\ 2009, The elements of statistical learning: data mining, inference, and prediction, Springer series in statistics (New York: Springer)

2009
[50]

Hertzsprung, E.\ 1926, , 3, 115

1926
[51]

T., Irwin, M

Hodgkin, S. T., Irwin, M. J., Hewett, P. C., & Warren, S. J.\ 2009, , 394, 675

2009
[52]

G., Han, J

Hou, L. G., Han, J. L., & Shi, W. B.\ 2009, , 499, 473

2009
[53]

G., & Han, J

Hou, L. G., & Han, J. L.\ 2014, , 569, A125

2014
[54]

G., & Han, J

Hou, L. G., & Han, J. L.\ 2015, , 454, 626

2015
[55]

J., Hindman, J

Kerr, F. J., Hindman, J. V., & Gum, C. S.\ 1959, Australian Journal of Physics, 12, 270

1959
[56]

J.\ 1964, Ann

Huber, P. J.\ 1964, Ann. Math. Statist. 35, 73

1964
[57]

S., Babler, B

Indebetouw, R., Mathis, J. S., Babler, B. L., et al.\ 2005, , 619, 931

2005
[58]

Inno, L., Matsunaga, N., Romaniello, M., et al.\ 2015, , 576, A30

2015
[59]

J., Lewis, J., Hodgkin, S., et al.\ 2004, , 5493, 411

Irwin, M. J., Lewis, J., Hodgkin, S., et al.\ 2004, , 5493, 411

2004
[60]

Ita, Y., Matsunaga, N., Tanab \'e , T., et al.\ 2018, , 481, 4206

2018
[61]

Ioffe, S., & Szegedy, C.\ 2015, arXiv e-prints , arXiv:1502.03167

2015 arXiv
[62]

S., Stanek, K

Jayasinghe, T., Kochanek, C. S., Stanek, K. Z., et al.\ 2018, , 477, 3145

2018
[63]

Jones, E., Oliphant, E., Peterson, P., et al.\ 2001, SciPy: Open Source Scientific Tools for Python, 2001-, http://www.scipy.org/

2001
[64]

Kalberla, P. M. W., Dedes, L., Kerp, J., & Haud, U.\ 2007, , 469, 511

2007
[65]

Kalberla, P. M. W., & Kerp, J.\ 2009, , 47, 27

2009
[66]

Kim, D.-W., & Bailer-Jones, C. A. L.\ 2016, , 587, A18

2016
[67]

P., & Ba, J.\ 2014, arXiv e-prints , arXiv:1412.6980

Kingma, D. P., & Ba, J.\ 2014, arXiv e-prints , arXiv:1412.6980

2014 arXiv
[68]

Koo, B.-C., Park, G., Kim, W.-T., et al.\ 2017, , 129, 094102

2017
[69]

E.\ 2012, Advances in Neural Information Processing Systems 25, 1097 (Curran Associates, Inc.)

Krizhevsky, A., Sutskever, I., Hinton, G. E.\ 2012, Advances in Neural Information Processing Systems 25, 1097 (Curran Associates, Inc.)

2012
[70]

Lecun, Y., Bottou, L., Bengio Y., Haffner, P.\ 1998, Proceedings of the IEEE, 86, 2278

1998
[71]

D., & Stobie, R

Laney, C. D., & Stobie, R. S.\ 1992, , 93, 93

1992
[72]

LeCun, Y., Bengio, Y.,\ 1995, The Handbook of Brain Theory and Neural Networks, 255 (MIT Press)

1995
[73]

S., Blitz, L., & Heiles, C.\ 2006, Science, 312, 1773

Levine, E. S., Blitz, L., & Heiles, C.\ 2006, Science, 312, 1773

2006
[74]

C., Yuan, C., & Shu, F

Lin, C. C., Yuan, C., & Shu, F. H.\ 1969, , 155, 721

1969
[75]

Lin, M., Chen, Q., & Yan, S.\ 2013, arXiv e-prints , arXiv:1312.4400

2013 arXiv
[76]

E.\ 2018, , 156, 171

Luck, R. E.\ 2018, , 156, 171

2018
[77]

M., Ngeow, C.-C., Kanbur, S

Macri, L. M., Ngeow, C.-C., Kanbur, S. M., Mahzooni, S., & Smitka, M. T.\ 2015, , 149, 117

2015
[78]

M., Ngeow, C.-C., Kanbur, S

Macri, L. M., Ngeow, C.-C., Kanbur, S. M., Mahzooni, S., & Smitka, M. T.\ 2016, , 151, 48

2016
[79]

J., Turner, D

Majaess, D. J., Turner, D. G., & Lane, D. J.\ 2009, , 398, 263

2009
[80]

Majaess, D., Turner, D., D \'e k \'a ny, I., Minniti, D., & Gieren, W.\ 2016, , 593, A124

2016
[81]

Matsunaga, N., Fukushi, H., Nakada, Y., et al.\ 2006, , 370, 1979

2006
[82]

Matsunaga, N., Kawadu, T., Nishiyama, S., et al.\ 2011, , 477, 188

2011
[83]

W., Kawadu, T., et al.\ 2013, , 429, 385

Matsunaga, N., Feast, M. W., Kawadu, T., et al.\ 2013, , 429, 385

2013
[84]

W., Bono, G., et al.\ 2016, , 462, 414

Matsunaga, N., Feast, M. W., Bono, G., et al.\ 2016, , 462, 414

2016
[85]

W., Emerson, J

Minniti, D., Lucas, P. W., Emerson, J. P., et al.\ 2010, , 15, 433

2010
[86]

N., Borissova, J., Catelan, M., et al.\ 2019, , 482, 5567

Molina, C. N., Borissova, J., Catelan, M., et al.\ 2019, , 482, 5567

2019
[87]

J., & Pierce, M

Monson, A. J., & Pierce, M. J.\ 2011, , 193, 12

2011
[88]

Morris, M., & Serabyn, E.\ 1996, , 34, 645

1996
[89]

M., Gould, A., Fouqu \'e , P., et al.\ 2013, , 769, 88

Nataf, D. M., Gould, A., Fouqu \'e , P., et al.\ 2013, , 769, 88

2013
[90]

Nishiyama, S., Nagata, T., Kusakabe, N., et al.\ 2006, , 638, 839

2006
[91]

Nishiyama, S., Tamura, M., Hatano, H., et al.\ 2009, , 696, 1407

2009
[92]

Oliphant, T.E., 2006, A guide to NumPy, Trelgol Publishing USA

2006
[93]

A., Palafox, L., & Griffith, C

Pearson, K. A., Palafox, L., & Griffith, C. A.\ 2018, , 474, 478

2018
[94]

E., Madore, B

Persson, S. E., Madore, B. F., Krzemi \'n ski, W., et al.\ 2004, , 128, 2239

2004
[95]

Pedregosa, F., Varoquaux, G., Gramfort, A., et al.\ 2011, Journal of Machine Learning Research, 12, 2825

2011
[96]

Pietrukowicz, P., Koz owski, S., Skowron, J., et al.\ 2015, , 811, 113

2015
[97]

Pietrzy \'n ski, G., Graczyk, D., Gallenne, A., et al.\ 2019, , 567, 200

2019
[98]

Y., Haghpanahi, M., et al.\ 2017, arXiv e-prints , arXiv:1707.01836

Rajpurkar, P., Hannun, A. Y., Haghpanahi, M., et al.\ 2017, arXiv e-prints , arXiv:1707.01836

2017 arXiv
[99]

J., Menten, K

Reid, M. J., Menten, K. M., Zheng, X. W., et al.\ 2009, , 700, 137-148

2009
[100]

W., Starr, D

Richards, J. W., Starr, D. L., Butler, N. R., et al.\ 2011, , 733, 10

2011
[101]

W., Starr, D

Richards, J. W., Starr, D. L., Miller, A. A., et al.\ 2012, , 203, 32

2012
[102]

Ripepi, V., Molinaro, R., Musella, I., et al.\ 2019, , 625, A14

2019
[103]

Russeil, D.\ 2003, , 397, 133

2003
[104]

K., Hempel, M., Minniti, D., et al.\ 2012, , 537, A107

Saito, R. K., Hempel, M., Minniti, D., et al.\ 2012, , 537, A107

2012
[105]

W., Fonnesbeck, C.\ 2016, PeerJ Computer Science, 2, 55

Salvatier, J., Wiecki, T. W., Fonnesbeck, C.\ 2016, PeerJ Computer Science, 2, 55

2016
[106]

J., Dame, T

Sanna, A., Reid, M. J., Dame, T. M., et al.\ 2017, Science, 358, 227

2017
[107]

F., Meisner, A

Schlafly, E. F., Meisner, A. M., Stutz, A. M., et al.\ 2016, , 821, 78

2016
[108]

Schmidt, M.\ 1957, , 13, 247

1957
[109]

K., Pasquali, A., et al.\ 2018, , 478, 3590

Shabani, F., Grebel, E. K., Pasquali, A., et al.\ 2018, , 478, 3590

2018
[110]

J., & Vanderburg, A.\ 2018, , 155, 94

Shallue, C. J., & Vanderburg, A.\ 2018, , 155, 94

2018
[111]

Simonyan, K., & Zisserman, A.\ 2014, arXiv e-prints , arXiv:1409.1556

2014 arXiv
[112]

M., Skowron, J., Mr \'o z, P., et al.\ 2019, Science, 365, 478

Skowron, D. M., Skowron, J., Mr \'o z, P., et al.\ 2019, Science, 365, 478

2019
[113]

M., Skowron, J., Mr \'o z, P., et al.\ 2018, arXiv:1806.10653

Skowron, D. M., Skowron, J., Mr \'o z, P., et al.\ 2018, arXiv:1806.10653

2018 arXiv
[114]

F., Cutri, R

Skrutskie, M. F., Cutri, R. M., Stiening, R., et al.\ 2006, , 131, 1163

2006
[115]

C., Lucas, P

Smith, L. C., Lucas, P. W., Kurtev, R., et al.\ 2018, , 474, 1826

2018
[116]

Srivastava, N., Hinton, G., Krizhevsky, A., Sutskever, I., Salakhutdinov, R.\ 2014, Journal of Machine Learning Research, 15, 1929

2014
[117]

K., et al.\ 2015, , 65, 297

Soszy \'n ski, I., Udalski, A., Szyma \'n ski, M. K., et al.\ 2015, , 65, 297

2015
[118]

K., et al.\ 2017, , 67, 297

Soszy \'n ski, I., Udalski, A., Szyma \'n ski, M. K., et al.\ 2017, , 67, 297

2017
[119]

B.\ 2006, in Astronomical Data Analysis Software and Systems XV (ed

Taylor, M. B.\ 2006, in Astronomical Data Analysis Software and Systems XV (ed. C. Gabriel et al.), ASP Conference Series, 351, p. 666

2006
[120]

Udalski, A., Soszy \'n ski, I., Pietrukowicz, P., et al.\ 2018, , 68, 315

2018
[121]

P.\ 2015, , 450, 4277

Vall \'e e, J. P.\ 2015, , 450, 4277

2015
[122]

P.\ 2017, The Astronomical Review, 13, 113

Vall \'e e, J. P.\ 2017, The Astronomical Review, 13, 113

2017
[123]

W.\ 2014, , 788, L12

Wang, S., & Jiang, B. W.\ 2014, , 788, L12

2014
[124]

Wang, S., & Chen, X.\ 2019, , 877, 116

2019
[125]

L., Wieland, F., McAlary, C

Welch, D. L., Wieland, F., McAlary, C. W., et al.\ 1984, , 54, 547

1984
[126]

Westerhout, G.\ 1957, , 13, 201

1957
[127]

J., Zheng, X

Xu, Y., Reid, M. J., Zheng, X. W., & Menten, K. M.\ 2006, Science, 311, 54

2006
[128]

J., Reid, M

Xu, Y., Li, J. J., Reid, M. J., et al.\ 2013, , 769, 15

2013
[129]

Xu, Y., Reid, M., Dame, T., et al.\ 2016, Science Advances, 2, e1600878

2016
[130]

B., Reid, M

Xu, Y., Bian, S. B., Reid, M. J., et al.\ 2018, , 616, L15

2018
[131]

R., Indebetouw, R., et al.\ 2009, , 707, 510

Zasowski, G., Majewski, S. R., Indebetouw, R., et al.\ 2009, , 707, 510

2009
[132]

Zechmeister, M., & K \"u rster, M.\ 2009, , 496, 577

2009

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [1]

, " * write output.state after.block = add.period write newline

ENTRY address archivePrefix author booktitle chapter doi edition editor eprint howpublished institution journal key month number organization pages publisher school series title misctitle type volume year version url label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts ...

[2] [2]

write newline

" write newline "" before.all 'output.state := FUNCTION format.url url empty "" new.block "" url * "" * if FUNCTION format.eprint eprint empty "" archivePrefix empty "" archivePrefix "arXiv" = new.block " " eprint * " " * new.block " " eprint * " " * if if if FUNCTION format.doi doi empty "" " " doi * " " * if FUNCTION format.pid doi empty eprint empty ""...

[3] [4]

Abadi, M., Agarwal, A., Barham, P., et al.\ 2016, arXiv e-prints, arXiv:1603.04467

arXiv 2016

[4] [5]

Gravity Collaboration, Abuter, R., Amorim, A., et al.\ 2019, , 625, L10

2019

[5] [6]

Abuter, R., Amorim, A., Bauboeck, M., et al.\ 2019, arXiv e-prints , arXiv:1904.05721

arXiv 2019

[6] [7]

Anders, F., Khalatyan, A., Chiappini, C., et al.\ 2019, arXiv e-prints, arXiv:1904.11302

arXiv 2019

[7] [8]

I., Eyer, L., & Mowlavi, N.\ 2013, , 434, 2238

Anderson, R. I., Eyer, L., & Mowlavi, N.\ 2013, , 434, 2238

2013

[8] [9]

I., Saio, H., Ekstr \"o m, S., et al.\ 2016, , 591, A8

Anderson, R. I., Saio, H., Ekstr \"o m, S., et al.\ 2016, , 591, A8

2016