Astronomical News

Report on the ESO Workshop Satellites and Streams in Santiago

held at ESO Vitacura, Santiago, Chile, 13–17 April 2015

Andreas H. W. Küpper1 the Milky Way (MW) and M31 pose chal- Steffen Mieske2 lenges to our understanding of structure and star formation in the early Universe, and the feedback between baryons and 1 Department of Astronomy, Columbia dark matter. University, New York, USA 2 ESO But how well do we understand what it means to be a satellite of the MW or M31? Even in the era of high-precision Galactic satellites and tidal streams are cosmology, we are still uncertain about arguably the two most direct imprints the total masses of the two dominant of hierarchical structure formation in the ­galaxies in the Local Group, their assem- haloes of galaxies. At this ESO work- bly histories and the shape and extent of shop we sought to create the big picture their dark matter haloes — key aspects of the galactic accretion process, and for gaining a consistent picture of these shed light on the interplay between galaxies and their satellite systems in a ­satellites and streams in the Milky Way, ΛCDM context. On the contrary, the dis- Andromeda and beyond. The Scientific covery of transition objects at the star Organising Committee prepared a cluster–dwarf galaxy interface has made well-balanced programme with 60 talks things more complicated. It has blurred and 30 poster contributions, resulting the historical distinction between satellite in a meeting which was greatly enjoyed classes, putting in question our under- by the more than 110 participants standing of tidal transformation and the at the venue, and worldwide via Twitter census of small stellar systems. Figure 1. Conference poster showing the Milky Way (#SSS15). with its most prominent satellite, the Large Magellanic Cloud, over Santiago. The upper part of the poster Although these aspects of near-field shows a collection of tidal streams from the Vía Lactea ­cosmology have become more and more Cauda simulation. Introduction prominent in the age of surveys, there has not been a conference on both satel- by presentations on observations and Near-field cosmology has become lites and streams in over a decade. This modelling of tidal streams, and the final increasingly important over the past few five-day ESO workshop (see the work- part of the workshop was dedicated decades. While the current concordance shop poster, Figure 1) therefore met with to the star cluster–dwarf galaxy divide. cosmological model (Lambda Cold Dark a great demand for presentation slots, The grouping into these sessions was Matter [ΛCDM]) has been very successful and the registration had to close early of course not strict, and, as intended, in reproducing and predicting the proper- due to the overwhelming interest. many presenters pointed out important ties of the Universe on large scales, connections between satellites and ­several possible tensions have been iden- Based on the scope of the meeting, the streams. The workshop programme, with tified on small scales (≤ 1 Mpc). Issues week was divided into three parts. First, links to many of the presentations, can like the “missing satellite problem”, the satellites and the satellite systems of the “core/cusp problem”, “too big to fail”, Milky Way, M31 and other nearby galax- Figure 2. Conference photo taken in the garden of and detections of satellite discs around ies were discussed. This was followed the ESO Vitacura premises. Stephane Courteau Stephane

38 The Messenger 161 – September 2015 be accessed online1. The participants are topic at the end of the meeting. Collins hosts and making them interesting test- shown in the gardens of ESO Vitacura in furthermore pointed out that kinematic beds for ΛCDM. Else Starkenburg showed Figure 2. data may often be ambiguous as some results from her semi-analytical models of the satellites seem to have had recent of dwarf galaxies, which are particularly The meeting was opened by the ESO gravitational interactions and encounters, useful to test the physics inside the satel- Director General Tim de Zeeuw, giving an while others like Hercules and Willman 1 lites and gain intuition about their evolution. overview of the ESO facilities and an appear to be entirely out of equilibrium, exciting outlook for the field into the Euro- making mass determinations hard, if not pean Extremely Large Telescope era, impossible. Accretion of satellites when it will be possible to resolve stellar populations out to the Virgo Cluster. Finding new satellites and following up Accretion and disruption of satellites the discoveries with deeper photo­metry can give us important insights into the and/or spectroscopy has become a build-up of bulges and haloes of galaxies. Satellite systems sport. Yet only through this important Benjamin Hendricks demonstrated this exercise will we eventually get a complete approach for globular clusters (GCs) in Galactic satellites give us an account of census of substructure in the local Uni- Fornax, and Ryan Leaman shed light on low-mass substructures at the present verse. After seeing the great successes the accretion history of the Milky Way day. However, in order to put these satel- of PAndAS, we heard further results and using its GCs as tracers of dwarf galaxy lites into the context of structure forma- outlooks from ongoing and upcoming infalls. Similarly, tidal streams provide tion within the ΛCDM framework, their surveys with the Large Sky Area Multi- insights on the star formation histories of masses, and especially their dark matter Object Fibre Spectroscopic Telescope individual dwarf galaxies, as Thomas content, have to be understood. In the (LAMOST) by Jeff Carlin, on the Large de Boer demonstrated in the example of first talk of the satellite session, Jorge Binocular Telescope (by Giacomo Beccari) Sagittarius. Peñarrubia discussed the advances, prob- and the Magellan Telescope (Denija lems and challenges of mass modelling Crnojevic), all looking for more substruc- But larger infalling galaxies, or major of dwarf galaxies. Although modelling ture in the local Universe. mergers, could also trigger the formation might help to put constraints on the nature of new satellites — known as tidal dwarf of dark matter (density cores/cusps), he Eduardo Balbinot­ then presented the rich galaxies. Pavel Kroupa argued that these clearly called for more (kinematic) data to harvest from the Dark Energy Survey satellites could be long-lived, and pollute inform the models. In the spirit of the (DES) data release 1, giving us a detailed (or entirely make up) the satellite popula- meeting’s title, and like many other speak- look at the structure and formation his- tions of host galaxies. Kinematically cor- ers, he also presented an interesting new tory of the Large Magellanic Cloud (LMC), related satellite populations in the Milky idea, using the streams of satellites to as well as of the nine newly discovered Way and M31 halo could point towards break the core/cusp degeneracy. Calling satellites around the Magellanic Clouds. such a formation scenario. However, it a “diversity problem” rather than a Later in the meeting, Vasily Belokurov Pierre-Alain Duc argued that the forma- cusp/core problem, Chervin Laporte dem- talked more about the independent dis- tion of tidal dwarf galaxies through onstrated the large phenomenology of covery of these satellites in the DES data. ­mergers is less likely than expected, and dark matter in galactic satellites due to a Both groups have follow-ups on one that the objects thus formed do not possible re-growth of cusps in cored ­satellite, Reticulum II, confirming that it is resemble the satellites found in the local ­galaxies via accretion of dark matter sub- in fact a dwarf galaxy. Similarly, Erik Universe. From the numerical side, Sylvia haloes or other dwarf galaxies — further ­Tollerud presented his newest findings in Ploeckinger is developing the tools to complicated by baryonic effects on the the Galactic Arecibo L-band Feed Array perform full hydrodynamical simulations central density profiles of satellites. (GALFA) survey. The two newly discov- of tidal dwarf galaxies to study their long- ered, gas-rich, low-mass galaxies repre- term survivability. Michelle Collins then gave a recount of sent the progenitors of the dwarf galaxies the dozens of newly discovered satellites that we find around larger host galaxies. Marcel Pawlowski, Rodrigo Ibata and in the Milky Way and M31 haloes from Understanding this transformation from Noam Libeskind then gave detailed surveys like the Pan-Andromeda Archae- gas-rich to quenched galaxy is not descriptions of the co-rotating structures ological Survey (PAndAS; see Figure 3). straightforward, as Thorsten Lisker ex­­ around the Milky Way, Andromeda and In her presentation, a mass–size diagram plained and Rachael Beaton later on con- Centaurus A. The speakers pointed out (Figure 4) made its first appearance and firmed. that the chances of finding such planes in ended up being the plot shown most ΛCDM appears to be rather low, but may often during the meeting. In particular the Numerical simulations are getting up have to do with either larger mergers divide between dwarf galaxies and glob- to speed with the flood of observational bringing in lots of satellites or infall of satel- ular clusters in this diagram is currently data. Coral Wheeler presented high-­ lites along the cosmic web. Gurtina Besla being populated with newly discovered resolution hydrodynamical simulations added the Large and Small Magellanic objects, most of which cannot be clas­ of isolated dwarf galaxies to understand Clouds to this picture, discussing different sified by their photometry alone — which their satellites, finding a few ultra-faint infall scenarios for the two satellite galax- is why we had an entire session on this satellites around each of these low-mass ies. She argued that if the LMC originally

The Messenger 161 – September 2015 39 Astronomical News Küpper A. & Mieske S., Report on the Workshop “Satellites and Streams in Santiago”

11 had a mass of > 10 MA, it should have brought in at least seven massive satel- lites. Moreover, she pointed out that a massive LMC may have shifted the bary­ 5 centre of the Milky Way, which could potentially affect all stream modelling. The satellite session thus concluded with lots of intriguing problems and open 0

questions. In many cases, satellite orbits ) appear to be the missing ingredients to answer these questions, which is where

tidal streams may come in handy. (Degrees

η –5

Tidal streams

Tidal streams are tracers of how satellites get accreted and disrupted. They make –10 up an as-yet unknown percentage of halo stars, complicating the modelling and interpretation of the stellar halo. But, due to their coherence in phase space, they 10 5 0 –5 –10 also enable us to measure the shape of ξ (Degrees) the gravitational potentials of their host galaxies, constrain the orbits of their pro- Figure 3. The surface density of colour-selected in the Halo of the Milky Way (most of genitors, and provide insights into the stars in the PAndAS footprint is shown. Andromeda which he discovered himself). Based on and M33 are clearly visible, as well as lots of smaller chemo-dynamical evolution of satellites. satellites and a complex network of streams. From the present-day globular cluster popula- As a first speaker of the session, Rodrigo Richardson et al. (2011). tion, Mark Gieles estimated that this Ibata presented modelling approaches number should be a factor of four times and promising streams that need to be higher, whereas Grillmair estimates it to modelled, such as the Giant Southern help to inform models of this system. Yet, be a factor of ten higher. Grillmair pointed Stream (GSS) in M31 (Figure 3). Karrie most insights on streams have come, out that, although the northern sky looks Gilbert then showed results from the and are still coming, from wide-field imag- like a “weaved carpet”, there is a sig­ Spectroscopic and Photometric Land- ing surveys. In this context, Blair Conn nificant lack of stream detections on the scape of Andromeda’s Stellar Halo presented results from the PanSTARRS1 southern hemisphere. (SPLASH), shedding light on the merger survey, concluding that the Monoceros event that created the GSS and related overdensity could be either a stream, a Aaron Romanowsky extended this cen- shells in the M31 halo. She pointed out flare or a ripple in the Galactic Disc. Heidi sus to M31 and beyond, emphasising that the halo of Andromeda shows clear Newberg argued for the latter, backing that streams are found in basically any signatures of further mergers with smaller up her arguments with signatures of such galaxy with deep enough photometry. He galaxies. Edouard Bernard added a density wave in Sloan Digital Sky pointed out that for distant galaxies, insights on the GSS’s star formation his- ­Survey (SDSS) data. A different approach since stars are too faint for spectra, glob- tory from Hubble Space Telescope (HST) of finding and characterising halo sub- ular clusters or planetary nebulae have data, and possible links to a star forma- structure was presented by Kathy Vivas, to be used to get kinematic data for inves- tion event in M31. who showed the extent of the Virgo over- tigating stream-like substructures. With density as seen by RR Lyrae stars. the example of M87, he then demon- Essential for our understanding of satel- strated that these tracers can in fact be lites, streams and their host galaxies are Putting all this into context, the “old cow” used to detect accretion events. Michael proper motion measurements. Tony Sohn of the stream business (his words), West followed up on that with the investi- presented HST observations, giving valu- ­Steven Majewski, gave an excellent his- gation of halo substructure using com- able insights on the internal kinematics, torical overview of how tidal streams were puter-vision techniques, that is, by apply- the gravitational potential of the Milky Way, discovered and modelled in the past ing the versatile Hough transformation to and even the kinematics of streams. twenty (forty?) years, and on how the dis- globular cluster maps. Iskren Georgiev presented first-epoch covery of the Sagittarius stream led to data from a Palomar 5 proper motion a paradigm shift in the astronomical com- measurement with the HST. The data munity, since it was a striking confirma- Modelling tidal streams provide tight constraints on the distances tion of hierarchical structure formation. to the cluster and its stream and their Carl Grillmair followed up with a complete Fast and efficient ways of modelling tidal stellar mass functions, which significantly census of the 21 currently known streams streams have become available in recent

40 The Messenger 161 – September 2015 Figure 4. This mass–size plot from $R C2OGR K,22VHSGm With this outlook we ended the streams Norris et al. (2014) depicts absolute 3+ 2# B$R -TBKDH 8,"R session and moved on to satellite proper- magnitude, as a tracer of stellar mass, C$RC2R K,22VN &"R4"#R , m versus effective radius, as a proxy for ties and tidal transformation. From the size. It contains nearly all known satel- variety of tidal features around satellites lites in the local Universe. Globular seen in the stream session it became  clusters (red) and dwarf galaxies (grey clear that we have to expect many satel- squares) form two distinct sequences, which merge onto a common mass– lites to be tidally affected, altering proper- size sequence defined by elliptical gal- ties like mass, size and velocity dispersion.  axies. Newly found objects close this gap between the two satellite popula-  tions. OB

 The star cluster–dwarf galaxy interface D

1  Properties of satellites differ substantially, from compact, relatively low-mass globu-  lar clusters to extended, high-mass dwarf galaxies with complex chemical compo- sitions. The discovery of transition  objects at the star cluster–dwarf galaxy l l l ll interface has blurred the his­torical dis-

,5L@F tinction between these satellite classes. Exploring this interesting region in the years. Yet, applications to real data are her modelling results, showing that the mass–size plane was therefore an inte- still the exception. With the flood of kine- only triaxial halo model currently known gral part of this workshop. matic data in mind, which will soon come for the Milky Way dark matter profile that from the Gaia satellite (mid 2016/early can successfully reproduce the Sagittarius One of the main questions was the nature 2017), Amina Helmi gave an excellent stream, fails in reproducing the thin and of ultra-compact dwarf galaxies (UCDs). overview of action-angle modelling of curved morphology of the Palomar 5 Are they massive globular clusters, com- streams and phase-space substructure. stream. She explained that the triaxiality pact dwarf galaxies or disrupting satel- She emphasised that streams are signifi- of the gravitational potential leads to sig- lites? In an attempt to answer this ques- cantly simpler structures in action-angle nificant stream fanning, resulting in puffy tion, Carolin Wittman presented imaging space than in phase space, which makes streams that do not resemble the SDSS data to search for tidal features around this for­malism a promising technique to observations of Palomar 5. Andreas low-mass galaxies in galaxy clusters. She model large numbers of streams. Based ­Küpper followed up on Pearson’s talk, by found signs of tidal disruption around one on the Aquarius simulations, she esti- identifying (probable epicyclic) substruc- UCD. Similarly, Michael West presented mated that Gaia should be able to detect tures within the stream of Palomar 5. He data on the M31 satellite G1, which shows about 400 streams in the Solar Neigh- demonstrated how these, when ­modelled clear signs of tidal disruption, and G1 bourhood. Building on Helmi’s talk, Jo correctly, turn globular cluster streams may therefore be a nucleated dwarf gal- Bovy introduced his Python galpy soft- into high precision scales, ­constraining axy rather than a globular cluster. ware, which can be used to efficiently models of Palomar 5, the Milky Way and generate stream models, and which is the Solar position and motion within the Oleg Gnedin showed that globular clus- entirely based on the simple nature of Galaxy. ters are fundamentally different from streams in action-angle space. Then dwarf galaxies. These dense systems Robyn Sanderson demonstrated how this But substructure in tidal streams may formed during galaxy mergers — from framework can be used to disentangle also arise from dynamical encounters the early gas-rich Universe to the present stream memberships of stars in the with dark matter subhaloes. Denis Erkal day — following a universal log-normal (messy) Galactic Halo, and in this way explained within a mathematical frame- mass function. Dissolving GCs in the constrain the potential of the Milky Way. work how these dark satellites affect inner regions of galaxies may spiral into ­stellar streams and what we can learn the centres and form nuclear star clus- The kinematically cold streams detected about dark matter from a single gap in a ters; dissolving GCs in the outer regions in the Milky Way Halo, such as GD-1 and stream. This will certainly be an important of galaxies, such as the Milky Way satel- Palomar 5, are very thin — surprisingly use for tidal streams in the future, since lite Crater (the true nature of which is thin. In this context, Raymond Carlberg ΛCDM predicts thousands of satellites actually still under debate), may appear demonstrated the effects of a realistic, around Milky-Way-size galaxies, whereas like ultra-faint dwarf galaxies. Relying on i.e., triaxial and substructured, dark mat- our current number count of luminous results from numerical simulations,­ Mike ter halo on the width of tidal streams. He satellites lingers around 40–50. Why Fellhauer then added that GCs, when concluded that the potential of the Milky some satellites contain luminous matter brought completely out of virial equilib- Way should rather be filled with streams and why most others are probably devoid rium through tidal shocks, can mimic ultra- looking like “overcooked spaghetti”. In of baryons has to be understood, if the faint dwarf galaxies (UFDs) like Hercules this context, Sarah Pearson presented ΛCDM model of cosmology is to prevail. and Segue 1. In a similar fashion, but with

The Messenger 161 – September 2015 41 Astronomical News Küpper A. & Mieske S., Report on the Workshop “Satellites and Streams in Santiago”

opposite sign, Filippo Contenta pre- strengthening the accretion scenario. It the densest stellar systems in the local sented results from N-body simulations was also shown that, even though new Universe. M60-UCD1 and M59-UCD3, showing that dissolving GCs are unlikely discoveries of extended clusters have being typical examples of these hyper- to contribute to the UFD population.­ But slowly filled up the mass–size plane, the compact clusters, even showing signs of not all star clusters that fall into in the two dominant populations of satellites central supermassive black holes. Anil gap between the bulk of GCs and dwarf are still “regular”­ GCs and dwarf galaxies. Seth presented integral field unit observa- galaxies within the mass–size plane (Fig- Many more of these exotic, extended tions of M60-UCD1, showing that this ure 4) are necessarily ultra-faint. Extended systems have been found outside the massive UCD indeed hosts a super­ outer-halo clusters like Palomar 14 or Local Group by the SAGES Legacy Uni­ massive black hole that makes up about Crater have half-light radii that are a fac- fying Globulars and GalaxieS (SLUGGS) 15 % of the galaxy’s mass. Such overly tor of ten times higher than usual for GCs. survey. massive black holes could be explained Paolo Bianchini weighed the hypothesis through tidal stripping of larger dwarf that these extended clusters could have ­galaxies in a host galaxy potential. been formed as ­compact within the Ultra-compact dwarf galaxies weaker tidal fields of dwarf galaxies, and In further presentations on the kinematics then fallen into the Milky Way Halo. He UCDs were, again, the topic of the final of GCs, UCDs, and compact ellipticals found that such clusters do indeed day of the meeting. Even after the previous (cEs), Mark Norris and Adrien Gouérou expand, but not enough to resemble the sessions, their origin was still unclear: demonstrated how important integral field extended GCs in the Milky Way, leaving are they the massive end of the globular units have become for the investigation of their origin as an open question. cluster mass function, or are they stripped extragalactic compact stellar systems, as nuclei of dwarf galaxies? The session they should all be called. In the final pres- What exactly is a galaxy? Several speak- was a back-and-forth of good arguments entation of the meeting, Florent Renaud ers tried to answer this controversial for each hypothesis. Michael Hilker gave a fascinating outlook on how the question. Jay Strader defined it as a pointed out that nuclear star clusters and formation of UCDs and GCs can be traced bound collection of stars whose proper- UCDs fall into the same region of the in high-resolution numerical simulations. ties cannot be explained by a combi­ mass–size plane. He also showed obser- He demonstrated how compact stellar nation of baryons and Newton’s law of vations of UCDs with clear signs of tidal systems can expand during galaxy merg- gravity. Referring to previous speakers, disruption. Yet he concluded that the ers, and form UCD-like objects. The he pointed out that a mass-to-light ratio number and properties of UCDs seem to meeting was concluded by the honorary is not a well-defined quantity for systems be incompatible with them being entirely conference photographer Stephane that are not in virial equilibrium, and, made up of stripped nuclei. In an attempt Courteau, who gave an entertaining sum- hence, should not be used for this distinc- to resolve the cluster/galaxy question, mary in pictures (and kindly made his con- tion. Instead, he proposed to use metal- Matthias Frank put forward the provoca- ference photos available to the public2). licity spread as a diagnostic to distinguish tive definition that “a globular cluster is a star cluster from a galaxy. It was noted, something that is made up of globular Based on the success of this workshop, however, that the existence of a mass– cluster stars”. With this definition in mind, the organisers hope that there will metallicity relation detected in extragalac- he found that X-shooter spectra of UCDs be many more Satellites and Streams tic GC systems (the blue tilt) suggests put them on the same sequence with ­meetings in the future. metallicity spreads to be a natural feature GCs, when looking at them in the CN– of massive GCs. Thus, not even a metal- MgFe plane. With even more ESO data, licity spread may be a sufficient criterion Karina Voggel found, from a large sample Acknowledgements to define a galaxy. Strader suggested of > 100 Fornax UCDs imaged with the The organisers would like to thank Maria Eugenia that, until enough data exists, newly found FOcal Reducer and low dispersion Spec- Gómez, Paulina Jirón, the team of ESO IT and objects should not be named, since trograph (FORS), that their distribution ­General Services, and the entire Local Organising ­otherwise star clusters like Crater end up around Fornax follows the distribution of Committee. Heartfelt thanks go also to ESO’s Direc- having names following the Local Group the GCs. She finds an overabundance torate for Science for supporting this workshop. dwarf galaxy naming convention. of GCs in the vicinity of UCDs, which she interprets as support for the stripping References In a series of excellent presentations, scenario — even though less than 20 % Dougal Mackey, Duncan Forbes, Jean show signs of tidal disruption in the form Norris, M. et al. 2014, MNRAS, 443, 1151 Richardson, J. C. et al. 2011, ApJ, 732, 76 Brodie and Mark Norris reviewed how of an extended stellar halo. Mark Norris, the mass–size diagram has filled up over furthermore, reasoned that UCDs with Links the last few decades (or rather centuries an absolute magnitude MV < 13 have to since the first discovery of a GC in 1665). be stripped nuclei, since they lie 5 away σ 1 Workshop programme with access to presentations: Many of these new objects have been from the cluster luminosity function. http://www.eso.org/sci/meetings/2015/Satel- found in M31, and most of those through lites2015/program.html the PAndAS survey. These data also No matter what they are, UCDs are 2 Gallery of conference photos: http://www.astro. show that the GCs in the outer halo of among the most extreme stellar systems. queensu.ca/people/Stephane_Courteau/gallery.php M31 are highly correlated with streams, Jean Brodie presented observations of

42 The Messenger 161 – September 2015