Surveys DOI: 10.18727/0722-6691/5128

4MOST Consortium Survey 9: One Thousand and One Magellanic Fields (1001MC)

Maria-Rosa L. Cioni 1 et al., 2013). This has motivated many MAgellanic Stellar ­History (SMASH), and Jesper Storm 1 studies aimed at explaining the structure the Optical Gravitational Lensing Experi- Cameron P. M. Bell 1 and the star formation history of the ment (OGLE), were made possible by the Bertrand Lemasle 2 Large Magellanic Cloud (LMC), the Small development of wide-field cameras at Florian Niederhofer 1 Magellanic Cloud (SMC) and their tidal ­telescopes dedicated to survey observa- Joachim M. Bestenlehner 3 features, i.e., the Magellanic Bridge and tions for a large fraction of the available Dalal El Youssoufi 1 Stream. Furthermore, with the increased time. Sofia Feltzing 4 number of deep imaging observations Carlos González-Fernández 5 we have discovered potential satellite Next to this wealth of photometric obser- Eva K. Grebel 2 ­galaxies of the Magellanic Clouds and vations, which have yet to reach their full David Hobbs 4 new stellar streams possibly associated exploitation (also including data from the Mike Irwin 5 with tidal stripping events (for example, Gaia satellite), there is a pronounced lack Pascale Jablonka 6 Koposov et al., 2018). of spectroscopic observations across Andreas Koch2 the range of stellar populations and sub- Olivier Schnurr 1, 7 The Magellanic Clouds are the largest structures of the Magellanic Clouds. The Thomas Schmidt 1 and most massive satellite galaxies of largest samples of moderate resolution Matthias Steinmetz 1 the Milky Way. The LMC resembles a (at least R = 4000) spectra, suitable for ­spiral galaxy, with a rotating disc, an off- kinematics and metallicity measurements, centre bar, and a few spiral arms. Young, comprise about 9000 giant stars (for 1 Leibniz-Institut für Astrophysik Potsdam intermediate-age and old stars show example, Dobbie et al., 2014). Chemical (AIP), Germany ­different levels of substructures extending tagging, a powerful tool to discern the 2 Zentrum für Astronomie der Universität to large radii. The LMC hosts the most history of stellar populations, requires Heidelberg/Astronomisches Rechen-­ massive stars known today (for example, high-resolution (at least R = 20 000) spec- Institut, Germany Bestenlehner et al., 2014). The SMC is a tra and these only exist for about 4000 3 Physics and Astronomy, University of dwarf spheroidal galaxy, with a significant giant stars and 1300 early-type massive Sheffield, UK depth along the line of sight and a mor- stars, and for much smaller samples of 4 Lund Observatory, Lund University, phology shaped by tidal interactions (for other stellar types (for example, Nidever Sweden example, Niederhofer et al., 2018). The et al. 2019). Despite the major scientific 5 Institute of Astronomy, University of SMC formed half of its stellar mass prior advances of these programmes, where Cambridge, UK to an age of ~ 6 Gyr (for example, Rubele two thirds of the spectra refer to LMC 6 Laboratoire d’astrophysique, École et al., 2018). The Magellanic Bridge is stars and one third to SMC stars, they are ­Polytechnique Fédérale de Lausanne, the product of an LMC–SMC collision far from providing a comprehensive view Switzerland ~ 200 Myr ago; it is likely formed of SMC of a system where stars span the age 7 Cherenkov Telescope Array Observ­atory, material and it contains both gas and range of the Universe and that is strongly Bologna, Italy stars. The origin of the Magellanic Stream, shaped by dynamical interactions. which is made of both LMC and SMC gas (for example, Richter et al., 2013), depends The One Thousand and One Magellanic on the orbital history of the Magellanic Specific scientific goals Fields (1001MC) survey aims to measure Clouds and their one-to-many interactions the kinematics and elemental abun- with the Milky Way and with each other. The 1001MC survey aims are as follows: dances of many different stellar popu­ – To find and characterise kinematic and lations that sample the history of for­ Large amounts of telescope time have chemical patterns within the Magellanic mation and interaction of the Magellanic been invested in imaging the Magellanic Clouds system. Clouds. The survey will collect spectra Cloud stars, studying their distribution, – To study links between kinematics and of about half a million stars with G < 19.5 and measuring their ages, distances, chemical patterns as well as their spa- magnitudes (Vega) distributed over an and motions. A major ESO programme, tial distribution across different stellar area of about 1000 square degrees and that will provide targets for spectroscopic populations. will provide an invaluable dataset for follow-up studies, is the VISTA survey – To establish how the star formation his- a wide range of scientific applications. of the Magellanic Clouds system (VMC 1), tory and the dynamical evolution of the aimed at deriving the spatially resolved system are related to these patterns. star formation history and three-dimen- – To study the metallicity-dependent Scientific context sional geometry of the system. The VMC physical and wind properties of massive is the most sensitive high-spatial-resolution stars and their evolutionary stages. During the last decade, our view of survey of the Magellanic Clouds in the – To quantify the metallicity dependence the Magellanic Clouds has changed sig­ near-infrared to date. The VMC and other of key distance indicators. nificantly. These galaxies most likely contemporary surveys in the optical approached the Milky Way only a few Gyr domain, such as the SMC in Time: Evolu- A comprehensive study of the kinematics ago, rather than having orbited around it tion of a Prototype interacting late-type and chemistry of a large number of stars for a Hubble time (for example, Kallivayalil dwarf galaxy (STEP), the Survey of the at different evolutionary phases and with

54 The Messenger 175 – March 2019 a wide spatial distribution is needed to radial velocity variations (for example, solar to very metal-poor (D[Fe/H] > 2 dex), address these goals. Sana et al., 2013) and derive approxi- and uncertainties in the Fe abundance mate systemic velocities of variable stars of 0.2 dex or better are needed to distin- Line-of-sight (radial) velocities are one of using templates, for example in the case guish between different stellar popula- the fundamental components of motion of Cepheids, RR Lyrae stars, and long-­ tions. These can be derived from spectra to describe the internal kinematics of gal- period variables (for example, Nicholls et of individual stars, but for some faint axies from which the distribution of mass al., 2010). ­targets the spectra will be combined to is estimated. Radial velocities, together reach a minimum signal-to-noise ratio with proper motions, are necessary to The 1001MC survey data will also be (S/N) of 20 per Å within the spectral derive space velocities from which to used to quantify and map the dust regions above. We aim to reach similar infer orbital motions. The 1001MC survey absorption within the Magellanic Clouds accuracies in measuring the abundance will provide radial velocities that match using background galaxies. This is of other prominent elements. the accuracy of the tangential velocities achieved by comparing the rest-frame (proper motions) that are measured using spectra of galaxies with a reddening free A fraction of the 1001MC targets are astrometry from, for example, the VMC template where the adjustment of the ­variable stars, which means that they survey (Niederhofer et al., 2018) and continuum level will provide a measure change temperature across the pulsation Gaia. These are of the order of 2.5 km s –1 of the dust content along the line of sight cycle. To derive elemental abundances for an ensemble of stars (which corre- (for example, Dutra et al., 2001). Com- for ­Cepheids it is necessary to acquire sponds to 1% of the radial velocity of the pared to an ongoing study, based on the spectra before the temperature changes Magellanic Clouds and is a factor of 10 analysis of spectral energy distributions, significantly and therefore we plan to smaller than the internal motion), consid- spectra provide the redshifts of galaxies observe for 1 hour at any given epoch, erably improving our capability to spa- which are needed to scale the templates, this time being sufficient to reach the tially sample kinematical substructures reducing considerably the uncertainties required S/N. RR Lyrae stars, instead, are within the Magellanic Clouds. associated with a photometric determi­ faint and for them we will focus on kine- nation. We estimate that with 120 galax- matics. However, we will explore the pos- The iron abundance [Fe/H] is usually ies per square degree we would obtain sibility of also measuring average Fe used as a proxy for the metallicity of stars. a dust map with a spatial resolution of abundances by combining, for example, Age and metallicity of red giant branch 0.143 square degrees, directly compara- stars with similar metallicities as esti- stars are, however, degenerate in the ble to that of current star formation history mated from the Fourier decomposition of ­colour-magnitude diagrams. Moreover,­ studies (for example, Rubele et al., 2018). their light-curves. the dependence on metallicity of key dis- tance indicators (for example, the peri- od-luminosity relations of Cepheids, the Science requirements Target selection and survey area luminosity of red clump stars) is still not well assessed. The 1001MC survey will The 1001MC survey aims to reach accu- The 1001MC survey will cover an area of derive the metallicity of different ­stellar racies of ± 2 km s –1 for the radial veloci- about 1000 square degrees (Figure 1). populations, provide a metallicity map of ties of individual stars. This accuracy is This area comprises targets that trace the system as a function of age, and lift designed to match the accuracy of the the extent of different stellar populations some of the degeneracies that affect proper motion obtained with other facili- and that describe substructures through- the tracers of stellar population parame- ties. For example, for individual bright out the Magellanic Clouds. The 1001MC ters. For bright stars, in addition to their stars in the Magellanic Clouds Gaia will survey will obtain 4MOST spectra of about radial velocity and iron abundance, we provide proper motion accuracies of half a million stars with G < 19.5 magni- will measure the abundances of several 2.5 km s –1 or obtain similar accuracies tudes and produce a sample that is a a-elements, Fe-peak elements, and on bulk velocities from the average of factor of 20 larger than the largest sam- ­elements produced in the slow (s-) and about 200 G-type or 3000 M-type stars. ple of Magellanic Cloud stellar spectra rapid (r-) neutron-capture nucleosynthesis These values have been calculated assembled in the past. Spectra from the processes (for example, Zr, Ba, Sr, Eu), assuming the Gaia end-of-mission proper Gaia Radial Velocity Spectrograph reach elements that will provide further con- motion accuracies 2 and considering that bright red and blue supergiant stars, straints on the chemical enrichment his- at the distance of the Magellanic Clouds while completed­ and ongoing observing tory of the different components of the 0.01 milliarcseconds per year is roughly campaigns with 2dF and APOGEE-2S two galaxies. We will also classify stars equivalent to 2.5 km s –1. observe giant stars well above the red throughout the Hertzsprung-Russell dia- clump in the Magellanic Clouds. The gram from their spectral features. In par- We also aim to obtain metallicities for 1001MC observations will prioritise areas ticular we will obtain a complete census individual stars with accuracies better with the highest legacy value (i.e., the of massive stars (> 15 M⊙) including O than ± 0.2 dex. This will be achieved central LMC and SMC areas). main-sequence stars, blue, yellow, and using Fe lines and/or indirectly using the red supergiants and Wolf-Rayet stars. Ca II and Mg b triplets. Stars in the The 1001MC targets span a range of about We aim also to provide some indication Magellanic Clouds span a relatively large 10 magnitudes (Table 1, Figure 2) and of the binary nature of these stars from range of metallicities, from about half comprise young massive and supergiant

The Messenger 175 – March 2019 55 Surveys Cioni M.-R. L. et al., 4MOST Consortium Survey 9

100 within the Magellanic Clouds. Table 1 10 N shows the approximate number of targets that belong to each sub-survey while 50 E ­Figure 2 shows the magnitude and colour 5 criteria adopted to select the targets.

30 N umber )

es 0 20 The central regions of the Magellanic re of

eg Clouds (a few hundred square degrees)

star will be observed more frequently (3–6 n (d 10 io –5 s visits) because of the high target density. pe

r This will then allow us to monitor a ­sub- pi xe Declin at –10 5 set of stars across the different types. Δ l Some stars will be monitored with the 3 HRS and others with the LRS. The main –15 goal of these monitoring campaigns is to 2 trace the variation in the radial velocity curves that are directly linked to the inter- –20 1 nal structure of pulsating stars and/or to 20 10 0–10 –20 the presence of companions, as well as ΔRight ascension (degrees) to establish the effect of binaries on the dynamics of the different stellar systems. Figure 1. Spatial source density distribution of the bles, and RR Lyrae stars). The number of In addition, we will make use of the poor 1001MC stars in a zenithal equidistant projection with targets reflects the need to statistically observing conditions programme (Guiglion the origin at a right ascension of 50 degrees and declination of –67.5 degrees. Pixels are 0.05 square characterise and spatially trace substruc- et al., p. 21) for those targets that fall degrees. The total area covered results from the tures of the Magellanic Clouds by age, within the 1001MC area. combination of two circular regions, each centred chemical content and, if possible, multi- on one of the Magellanic Clouds, that encompass ple diagnostics. their extended structure and possible tidal features. The circular cutout in the coverage at the bottom of Spectral success criteria the figure is caused by a declination > –80 degree The selection of 1001MC targets results selection limit, which was implemented because from the combination of near-infrared To meet the goals of the 1001MC survey observing at lower declinations becomes really observations from VISTA and 2MASS with we require spectra with a S/N per Å of ­inefficient with 4MOST. optical observations from Gaia, where about 100–1500 encompassing both LRS variable stars are identified in OGLE and and HRS observations of targets distrib- stars, intermediate-age giant stars Gaia data. In particular, Gaia parallaxes uted across the Hertzsprung-Russell dia- (asymptotic giant branch stars of M and are used to remove Milky Way stars and gram. Depending on stellar magnitude C type, red clump stars), old red giant to create a catalogue with homogeneous and type of star, we will measure elemen- and horizontal branch stars. They also coordinates. The VISTA selection data tal abundances and/or radial velocities. include different types of variable stars originate from the VMC survey for the In practice, for the abundance of specific (Cepheids of any type, long-period varia- central regions of the Magellanic Clouds elements we will use spectra with S/N > and from the VISTA Hemisphere Survey 80–100 per Å, while for determining indi- (VHS) for the outer regions. VISTA data rect Fe abundances with either the Ca II Table 1. Sub-surveys and preliminary numbers of will be used to select targets for the Low- and the Mg b triplets we will use spectra 1001MC targets. For the main sequence, red clump Resolution Spectrograph (LRS) while with S/N > 20–30 per Å. From all of the giant and red giant branch stars, these numbers ­targets for the High-Resolution Spectro- other spectra of individual stars with S/N represent only fractions of the respective popula- graph (HRS) will be selected from 2MASS. < 20 per Å we will derive radial velocities. tions (Figure 2; LRS only): 15% for main sequence stars, 5% for red clump stars, and 10% for red giant Several sub-­surveys are defined to repre- Furthermore, we will combine spectra branch stars. sent the range of stellar populations of individual stars, for example RR Lyrae stars, to estimate median element abun- Star type Spectrograph Magnitude range Number of targets dances and radial velocity. (103) Main sequence stars LRS 11.9 < G < 19.5 87 The S/N is measured in the continuum Red clump giant stars LRS 16.9 < G < 19.5 64 within one of the following spectral ranges: Red giant branch (RGB) stars LRS, HRS 14.1 < G < 19.5 310 5140–5200 Å (including the Mg b triplet RR Lyrae stars LRS 12.2 < G < 19.5 36 at: 5167, 5172, 5183 Å) and 8350–8850 Å Cepheids HRS, LRS 11.5 < G < 19.5 9 (including the Ca II triplet at: 8498, 8542, Supergiant stars HRS, LRS 10.5 < G < 19.5 36 and 8662 Å). In this way there will be at O-rich asymptotic giant branch (AGB) stars HRS 10.8 < G < 19.5 25 least one spectral region with sufficient Carbon stars HRS 12.4 < G < 19.5 9 S/N to derive the radial velocity and some Background galaxies LRS 120 elemental abundance(s) of each star.

56 The Messenger 175 – March 2019 6 11 Figure 2. Near-infrared, LMC J–Ks versus Ks, A: Supergiants colour-magnitude Hess 7 12 B: O-rich AGB stars diagrams of the 1001MC A C: RGB stars stars across the LMC 13 (targets in the SMC 8 D: Carbon stars B x : Cepheids are ~ 0.5 magnitudes 14 D fainter). The red dashed 9 lines indicate the sepa- D rations between the ) ) 15 HRS targets (left) and ag ag 10 A B the LRS ­targets (right), (m (m except for Cepheids. All

Ks Ks 16 sources have G < 19.5 11 magnitudes except 17 A C LMC for the RR Lyrae stars that extend to fainter 12 A: Main-sequence stars 18 B: Supergiants magnitudes. C C: Red clump giants 13 19 D: RGB stars : Cepheids : RR Lyrae stars 14 20 –0.5 0.00.5 1.01.5 2.02.5 –0.250.000.250.500.751.001.25 J–Ks (mag) J–Ks (mag)

To estimate the progress of the 1001MC Acknowledgements Koposov, S. et al. 2018, MNRAS, 479, 5343 survey we define a figure of merit (FoM). Kozlowski, S. et al. 2013, ApJ, 775, 92 We acknowledge funding from the European Nicholls, C. P. et al. 2010, MNRAS, 405, 1770 The FoM for each sub-survey is the ratio Research Council (ERC) under the European Nidever, D. et al. 2019, ApJ, submitted of observed targets to the goal number, Union’s Horizon 2020 research and innovation pro- Niederhofer, F. et al. 2018, A&A, 613, L8 and the FoM for the whole survey corre- gramme (grant agreement No 682115 as well as Richter, P. et al. 2013, ApJ, 772, 111 sponds to the minimum FoM among from the German Research Foundation (DFG) via Rubele, S. et al. 2018, MNRAS, 478, 501 Sonderforschungsbereich “The Milky Way System” Sana, H. et al. 2013, A&A, 550, A107 the sub-surveys. In this way each stellar (SFB 881). ­population will be sufficiently represented across the Magellanic Clouds system. Links References 1 VISTA survey of the Magellanic Clouds system Bestenlehner, J. M. et al. 2014, A&A, 570, A38 (VMC): http://star.herts.ac.uk/~mcioni/vmc/ Dobbie, P. D. et al. 2014, MNRAS, 442, 1663 2 Expected nominal science performance of the Gaia Kallivayalil, N. et al. 2013, ApJ, 764, 161 mission: https://www.cosmos.esa.int/web/gaia/ science-performance

This VISTA image of the 30 Doradus star-forming region (or Tarantula Nebula) from the VMC Public Survey is being used to study the detailed star formation history of the Magellanic

ESO/M.-R. Cioni/VISTA Magellanic Cloud survey. Acknowledgment: Cambridge Astronomical Survey Unit Survey Astronomical Cambridge Acknowledgment: survey. Cloud Magellanic Cioni/VISTA ESO/M.-R. system.

The Messenger 175 – March 2019 57