The Messenger 184

Impact of Satellite Constellations and ESO’s Role in Supporting the Community
2020 Scientific Prioritisation Community Poll

The INvestigate Stellar Population In RElics (INSPIRE) Project
ESO’s First e-Conference #H02020
The Messenger
No. 184 | 2021
ESO, the European Southern Observa- Contents
tory, is the foremost intergovernmental
astronomy organisation in Europe. It is The Organisation
supported by 16 Member States: Austria, Williams, A. et al. – Analysing the Impact of Satellite Constellations and
Belgium, the Czech Republic, Denmark, ESO’s Role in Supporting the Astronomy Community 3
France, Finland, Germany, Ireland, Italy, Mérand, A. et al. – Report on the Scientific Prioritisation Community Poll (2020) 8
the Netherlands, Poland, Portugal, Spain,
Sweden, Switzerland and the United Instrumentation
Kingdom, along with the host country of Zwaan, M. et al. – A Guide to ALMA Operations and Interactions with
Chile and with Australia as a Strategic the Community 16
Partner. ESO’s programme is focussed Quertier, B. et al. – Upgrade Strategies for the ALMA Digital System 20
on the design, construction and opera-
tion of powerful ground-based observing Astronomical Science
facilities. ESO operates three observato- Spiniello, C. et al. – The INvestigate Stellar Population In RElics (INSPIRE)
ries in Chile: at La Silla, at P
aranal, site of Project — Scientific Goals and Survey Design 26
the Very Large Telescope, and at Llano
de Chajnantor. ESO is the European Astronomical News
partner in the Atacama Large Millimeter/ Anderson, R. I. et al. – Maintaining Scientific Discourse During a
submillimeter Array (ALMA). Currently Global Pandemic: ESO’s First e-Conference #H02020 31
ESO is engaged in the construction of the Miles-Páez, P. A. & Gentile Fusillo, N. P. – Fellows at ESO 37
Extremely Large Telescope. Personnel Movements 39
The Messenger is published, in hardcopy

and electronic form, four times a year.
ESO produces and distributes a wide
variety of media connected to its activities.
For further information, including postal
subscription to The Messenger, contact
the ESO Department of Communication at:
ESO Headquarters
Karl-Schwarzschild-Straße 2
85748 Garching bei München, Germany
Phone +498932006-0
information@eso.org
The Messenger
Editor: Mariya Lyubenova
Editorial assistant: Isolde Kreutle
Copy-editing, Proofreading:
Peter Grimley
Layout, Typesetting, Graphics:
Mafalda Martins
Design, Production: Jutta Boxheimer
www.eso.org/messenger/
Printed by omb2 Print GmbH,

Lindberghstraße 17, 80939 Munich,
Germany
Unless otherwise indicated, all images in

The Messenger are courtesy of ESO,
except authored contributions which are Front cover: This photo, taken at ESO’s Paranal Observatory, shows
the four Unit Telescopes and one of the Auxiliary Telescopes of the
courtesy of the respective authors.
Very Large Telescope (VLT) and the VLT Survey Telescope (VST), cap-
tured from an unusual perspective. Taken using a fish-eye lens, this
© ESO 2021 photography technique produces a 360 degree view of the location —
ISSN 0722-6691 creating an immersive Paranal world with the swirling Milky Way at the
centre of it. Credit: ESO/B.Tafreshi (twanight.org)
2 The Messenger 184 | 2021

The Organisation DOI: 10.18727/0722-6691/5237
Analysing the Impact of Satellite Constellations and

ESO’s Role in Supporting the Astronomy Community
Andrew Williams 1 purpose, under shared control” (Wood, Constellation Altitude Number of
Olivier Hainaut 1 2003). Currently operating constellations (Registering nation) (km) satellites
Angel Otarola 1 serve a variety of important and crucial Starlink Generation 1
updated (US) 550 1584
Gie Han Tan 1 functions for society, including: navigation
540 1584
Giuliana Rotola 1 and geodesy (for example, GPS, Galileo
570 720
and GLONASS), satellite telephony (for
560 348
example, Iridium), internet and TV (for
1 560 172
ESO example, ViaSat, Orbcom, GlobalStar) and
Starlink Generation 1
Earth observation (for example, Copernicus Phase 2 (US) 335 2493
and Planet). In the future, companies such 341 2478
In the coming decade, up to 100 000 sat- as SpaceX, Amazon, Samsung, Telesat 346 2547
ellites in large constellations could be and OneWeb, and several national entities Starlink Generation 2 (US) 328 7178
launched into low Earth orbit. The satel- (for example, the Chinese and Indian 334 7178
lites will introduce a variety of negative Space Agencies) are planning very large 345 7178
impacts on astronomy observatories and constellations in low Earth orbit (LEO). 360 2000
science, which vary from negligible to These systems aim to provide low-latency 373 1998
very disruptive depending on the type broadband internet around the world to 499 4000
of instrument, the position of the sci- support the “Internet of Things” to connect 604 144
ence target, and the nature of the con- directly machines and systems, financial 614 324
stellation. Since the launch of the first and gaming transactions, and military OneWeb Phase 2 reduced
batch of SpaceX’s Starlink constellation applications. Their ultimate goal is to pro- (US. UK) 1200 1764
in 2019, the astronomy community has vide global high-bandwidth connectivity, 1200 2304
made substantial efforts to analyse the including to remote places such as the 1200 2304
problem and to engage with satellite middle of an ocean or a remote village Amazon Kuiper (US) 590 784
operators and government agencies. (Curzi, Modenini & Tortora, 2020). 610 1296
This article presents a short summary 630 1156
of the simulations of impacts on ESO’s We were asked by the ESO Director Guo Wang GW-A49 (China) 590 480
optical and infrared facilities and ALMA, General to analyse the impacts on ESO’s 600 2000
as well as the conducted observational facilities and to support the emergent 508 3600
campaigns to assess the brightness community efforts with national societies 1145 1728
of satellites. It also discusses several and the IAU to study the issue and iden- 1145 1728
activities to identify p olicy solutions at tify mitigations. In this article, we present 1145 1728
the international and national level. a short summary of the outcome of the 1145 1728
simulations of the impact on ESO’s optical Sat Revolution (Poland) 350 1024
and infrared facilities and the observa- CASC Hongyan (China) 1100 320
Introduction tional campaigns conducted, and discuss
CASIC Xingyun Lucky Star
several activities to develop policy solutions (China) 1000 156
The optical astronomy community was at the international and national level. CommSat (China) 600 800
taken largely by surprise upon the launch Xinwei (China) 600 32
of the first batch of SpaceX’s Starlink AstromeTech (India) 1400 600
satellite constellation on 23 May 2019. In Impacts on ESO facilities and astro- Boeing (US) 1030 2956
their immediate post-launch configuration nomical science LeoSat (Luxembourg) 1423 108
the 60 satellite units were visible as a very Samsung (Korea) 2000 4700
bright “string of pearls” travelling at high Visible and infrared spectral range Yaliny (Russia) 600 135
velocity across the night sky, and gener- Telesat LEO (Canada) 1000 117
ated substantial public, media and astron- Drawing on public filings to the Interna- Total 78 265
omy community interest. ESO, along with tional Telecommunications Union (ITU)
many other observatories, national agen- and also national regulatory agencies, we Table 1. Planned satellite constellations. This table
reflects publicly available filings for spectrum from
cies, societies and the International Astro- estimated the number of planned satellite
the International Telecommunication Union (ITU) and
nomical Union (IAU) issued public state- constellations and the numbers of their national communications regulators. Projects are
ments1. The community quickly became individual units, as shown in Table 1. The at varying stages of approval. The data for Starlink
aware of the plans of many other com following work uses the Starlink 1st and and OneWeb include recent changes filed at the US
Federal Communication Commission (FCC). Some of
panies and nations to develop similar 2nd generations, OneWeb and GuoWang
the operators have withdrawn their applications (for
constellations. as a representative worst case scenario, example, Boeing, LeoSat). Only Starlink and OneWeb
totalling over 60 000 satellites between have launched operational satellites (over 1737 and
A satellite constellation is defined as “a altitudes of 300 and 1200 kilometres. We 218, respectively, as of June 2021). Many more com-
panies have filed applications for other purposes
number of similar satellites, of a similar used three independent methods to eval-
such as remote sensing. As these are typically much
type and function, designed to be in simi- uate the number of satellite trails that will smaller — and hence fainter — than telecommunica-
lar, complementary, orbits for a shared cross a field of view, as a function of the tion satellites, they are not considered in the analysis.
The Messenger 184 | 2021 3

The Organisation Williams, A. et al., Analysing the Impact of Satellite Constellations
Observatory latitude: – 24.6° Constellation: Observatory latitude: – 24.6° Constellation:

SL1+2, OW2r, GW1+2 Instrument: Fors2 imaging SL1+2, OW2r, GW1+2
Total 61 290 satellites 5 Field of view: 6ಿ x 6ಿ Total 61 290 satellites
N 50
Exposure time: 300 s N
2 10
1 4
Number of satellites degree – 2

0.4 1
Fraction lost
0.2 0.3
E W E W
0.09 0.07
0.04 0.02
0.02 0.005
0.008 S 0.001
Sun: S Sun:
Local time: 19:46 Local time: 19:46 Vsat in (10.3, 4.6) Veff in (22.9, 17.8)
δ: 0.00° 0.004 δ: 0.00° Selection: all satellites, scalled for effect 0.0004
Elevation: – 24.00 Elevation: – 24.00 Detected: Veff < 26.9 Bleeding: Veff < 18.5
Figure 1. (Upper left) An example distribution of sat- the density of illuminated satellites (in objects Figure 2. (Upper right) The effect of the satellites on
ellites over Paranal observatory. This figure shows a degree – 2). The east-west features correspond to the an exposure. As in Figure1, this map of the sky
map of the sky above Paranal at the beginning of an edges of constellations whose inclinations are close above Paranal observatory shows the effect of the
equinox night, showing a representative distribution to the latitude of the site. The dark crescent to the satellites on a 300-second image obtained with the
of Starlink, OneWeb, and GuoWang satellites. Of the east marks the area where all satellites are already in FOcal Reducer and low dispersion Spectrograph
total of 61 290 satellites, 2562 are above the horizon, the Earth’s shadow. (FORS), as a function of the direction of pointing. The
including 1373 illuminated ones. The colour indicates high airmass regions are hatched. The colour indi-
cates the fraction of exposures lost.
instrument characteristics, the pointing the detector destroy the whole observa- nation, which could be difficult to disentan-
azimuth and altitude, the local time of tion. Figure 2 shows an example for a gle from the science signal. Vera C. Rubin
night, and the time of year. The methods 300-second image with FORS. The losses, Observatory, with its 3 degree imager
can either produce statistics for the whole averaged over zenithal distances below behind an 8-metre telescope, is particu-
or numerical values for specific directions 60 degrees (airmass 2 and better, repre- larly affected. Its situation is made worse
of observation, fields of view and integra- senting 95% of observations with the by the effect of saturated trails on the
tion times. The three methods give good Very Large Telescope), are presented in camera, resulting in losses of up to about
agreement; further technical details are Figure 3 for a series of representative 30% for observations at twilight.
available in Walker et al. (2020). Figure 1 instruments. Because of the apparent
displays the result of one of these simula- concentration of satellites towards the The situation in the thermal infrared is com-
tions. The effect on observations scales horizon, the values in the figure should pletely different: the satellites are emit-
with the number of satellites, so the effect be about doubled when setting the limit ters, so they remain a source of contami-
of the current ~ 2000 constellation satel- at a zenithal distance of 70 degrees, and nation even when in Earth’s shadow, and
lites is about 3.5% of what is discussed halved when considering only 30 degrees they are well above the detection thresh-
below. around the zenith. The figure shows that old of instruments like the VLT Imager and
most instruments will suffer losses at the Spectrometer for mid-InfraRed (VISIR)
To evaluate the losses to an observation, ≲ 1% level at twilight, that figure dropping (Hainaut & Williams, 2020). However, the
we counted the number of satellite trails by 1–2 orders of magnitude when the field of view of a thermal IR instrument
crossing the field of view during the expo- sun reaches an elevation of – 40 degrees. is small, and the individual exposure time
sure, considering the brightness and Because of their brighter limiting magni- is very short, so only a negligible fraction
apparent angular velocity of each satellite. tudes, the high-resolution spectrographs of the exposures will be affected. A satel-
We excluded those that are too faint to are essentially immune to the satellites lite passing in front of a star will cause a
be detected. Those that are detected considered. The low-resolution spectro- short eclipse. However, because of their
destroy a 5-arcsecond-wide trail across graph will be able to register some of the high apparent angular velocity, it lasts of
the observation, and those saturating satellites as low-signal-to-noise contami- the order of 1 millisecond, and will cause

the brightness of satellites. SpaceX trialled
10 0 a dark coating on “DarkSat” Starlink,
which achieved some brightness reduc-
Midnight summer
Midnight equinox
LSST
Midnight winter
tion, but had implications for the thermal
control and was subsequently abandoned
10 –1
in favour of attitude adjustments and a
new “VisorSat2”. This features the addition
of a sunshade which protects the body of
4MOST FORSspec
Fraction lost
10 – 2 the satellite from direct illumination, and

MICADO FORSimg
which is mounted so that the illuminated
VST side of the shade is not visible from Earth.
Preliminary measurements suggest dark-
10 – 3 ening by a factor of about 3 compared to
HAWK-I
first-generation satellites. VisorSats have
been launched since mid-2020. Adjust-
Astronomical
ESPRESSO
Nautical
ments of the attitude of the satellites have

Sunset
10 – 4
Civil
Imager also reduced their brightness: the orien-

UVES
Low-res spectrograph tation of the solar panel has been adjusted

V V
High-res spectrograph to keep it invisible from the ground, hidden
10 – 5 behind the bus of the satellite.
0 –18 – 36 – 54 – 72 – 90
Sun elevation (degrees)
Figure 3. The mean fraction of exposure lost due to Stable Spectroscopic Observations (ESPRESSO) — Millimetre and submillimetre spectral
satellite trails as a function of the solar elevation, for 1 hour. The simulations account for the limiting mag-
range
representative exposures obtained at zenithal dis- nitude and saturation magnitude of the systems; they
tances smaller than 60 degrees. The twilights and assume that a detected satellite ruins a 5-arcsecond-
the solar elevation at midnight are indicated with wide trail across the detector, and that a saturated Of the satellite constellations listed in
shading. The exposures are as follows: the Vera C. satellite ruins the whole exposure. In the case of Table 1 the Starlink Generation 1 may
Rubin Observatory LSST Camera — 15 seconds; fibre-fed spectrographs (ESPRESSO, 4MOST), a
impact ALMA as downlink transmitters
the 4-metre Multi-Object Spectrograph Telescope detected satellite ruins the exposure. For 4MOST, a
(4MOST) on the Visible and Infrared Survey Tele- multi-fibre spectrograph, the number of trails in the since the end-user terminals are planned
scope for Astronomy (VISTA) — fraction of contami- field is converted into the number of affected fibres to operate at frequencies between 37.5
nated fibres for 1200 seconds; the VLT Survey Tele- using a Monte-Carlo simulation which indicates that and 42.5 GHz, which is within the ALMA
scope (VST) — 300 seconds; FORS — 1200 seconds on average a satellite will affect 1.254 fibres. FORS
Band 1 observing range. Whilst operator
for spectroscopy, 300 seconds for imaging; the images are affected at a level comparable to that of
High Acuity Wide-field K-band Imager (HAWK-I) — the much wider field VST because it is sensitive to agreements and observation schedul-
60 seconds; the Multi-AO Imaging Camera for Deep much fainter satellites. In the cases of ESPRESSO ing could avoid the unlikely case of direct
Observations (MICADO) — 60 seconds; the Ultravio- and UVES, the spectrographs are not sensitive to illumination of the ALMA antennas, the
let and Visual Echelle Spectrograph (UVES) — 1 hour, the fast-moving satellites.
cumulative background noise will be a
the Echelle SPectrograph for Rocky Exoplanet and
more persistent problem. We assume
that the interference from each satellite
a detectable dip in the signal of the star ure 2), or consider shutter-control meas- adds in an uncorrelated way in an ALMA
only for very fast photometry (< 1 second). ures and additional small telescopes, or receiver, meaning that the signal power
suitable wide-field cameras, to detect of each satellite can just be added to
Several mitigations have been identified satellites during or just before observa- obtain the total power received from all
that can reduce the impacts on astro- tions so as to either avoid them or identify satellites. Under these assumptions a
nomical facilities. The first option would be contaminated data. The level and com- total noise increase due to the Starlink
to reduce the number of satellites, lower plexity of the telescope or operations Generation 1 satellites of about 50 milli
their orbital altitude and minimise their mitigations put in place must be com- Kelvins at the input of an ALMA Band 1
size. Existing constellation designs, how- mensurate to the effect they correct. For receiver is expected. Longer integration
ever, are already highly optimised and many types of observations, it might be times should compensate for this noise.
linked to the spectrum assigned by regu- cheaper and more efficient just to repeat Where this cannot be done, reductions
latory authorities. Post-hoc changes are a failed observation. Currently, the impact in sensitivity will have to be accepted,
likely to be challenging, and therefore the on ESO visible and infrared facilities will although this requires more detailed study.
community should focus on influencing remain low enough that no telescope/ If a significant noise increase is expected,
operators in the early design stages. instrument or scheduling mitigations are this should be included in the ALMA
Other operator mitigations focus on design foreseen, but we are closely monitoring Sensitivity Calculator so that the correct
measures to darken satellites and avoid developments. integration times can be derived. Further
reflected sunlight. Observatories can details are available in a report made for
introduce scheduling based on the simu- The astronomy community, particularly in ESO’s 155th Council Meeting3.
lation results (for example, pointing to the US, has had a very productive collab-
the darkest areas in maps like that of Fig- oration with SpaceX aimed at reducing

The Organisation Williams, A. et al., Analysing the Impact of Satellite Constellations
The Starlink Generation 2 constellation make it less reflective of sunlight and Policy activities and the way forward
does not mention the use of a downlink effectively dimmer.
to user terminals in ALMA Band 1; this The simulations and observations outlined
function is foreseen at lower frequencies The satellites Starlink 1130 (DarkSat) and here have been facilitated by the coordi-
where ALMA does not observe. However, Starlink 1113b, launched on 7 January nated efforts of several national societies,
a request has been made as part of the 2020, were observed on 3 March 2020 such as the American Astronomical
US Federal Communication Commission immediately after arriving at their nominal Society, the Royal Astronomical Society
(FCC) licence application to operate a operational orbital height of 550 kilo- and the European Astronomical Society,
gateway downlink in the frequency range metres. The analysis of the observations amongst others, along with input from
71–76 GHz which falls in ALMA Band 2. produced a brightness magnitude (in observatories such as ESO, Vera C. Rubin
Since the number of gateways will be Sloan Digital Sky Survey g’ filter, scaled Observatory, NSF’s NOIRLab and the
much less than the number of end-user to zenith and range 550 kilometres) of Square Kilometre Array Observatory. This
terminals, interference to ALMA is expected 5.33 ± 0.05 and 6.10 ± 0.04 magnitudes body of work has established a basis for
to be a lesser concern. This assumes that for Starlink 1113 and 1130 (DarkSat), understanding the problem space and
these gateway terminals are not located respectively (Tregloan-Reed et al., 2020). yielded promising early results on opera-
close to the ALMA observatory. Planning These results showed that a) the reduc- tor mitigations. The bilateral work with
the location of the gateway terminals tion in brightness for DarkSat was 50.8% companies is welcome but is not a sus-
needs licensing from the Chilean national ± 3.5%, b) the brightness magnitude of tainable solution as the number of private
authorities and it is recommended that DarkSat (6.10 magnitudes) was still within and public constellation projects around
the Joint ALMA Observatory be actively the range of the naked-eye limiting mag- the world grows. Astronomers are now
involved in this process. At the time of nitude of an experienced observer (Bortle, beginning to look for regulatory solutions
writing, SpaceX has submitted a request 2001) and c) DarkSat was still brighter at both international and national levels.
to the Chilean regulatory body, SubTel, to than the limiting magnitude of 7 that has
operate four gateways on Chilean territory. been recommended to minimise unwanted At the international level, ESO supported
effects in sensitive astronomical cameras the IAU in submitting a set of policy recom-
(Walker et al., 2020; IAU 2020). Subse- mendations4, developed in the Dark and
Satellite observations quent observations of these satellites were Quiet Skies project (IAU 2020), to the
conducted in various spectral bands to United Nations Committee on the Peace-
Satellite observations have provided an assess their brightness and the effec ful Uses of Outer Space (COPUOS), which
important complement to simulations tiveness of the darkening treatment used addresses international dialogue on gov-
to calibrate assumptions and to charac- in DarkSat, in various astronomical spec- ernance of the use and exploration of
terise brightness over a range of wave- tral bands from visible to near-infrared outer space. The recommendations call
lengths and from different locations and (Tregloan-Reed et al., 2021). The results for, inter alia, international agreement to
orbital geometries. Satellites exhibit highly of these observations, conducted using address impacts on astronomy by mini-
dynamic changes in apparent brightness the Chakana telescope (Char et al., 2016) mising the orbital altitude, the brightness
that are due to geometrical and opera- and VIRCAM (Emerson, McPherson & of space objects to less than unaided-
tional factors, requiring many observations Sutherland, 2006; Dalton et al., 2006) on eye levels (7 magnitude), and antenna
to capture the changes in brightness over the VISTA telescope, are summarised in sidelobe emissions such that their indirect
the full range of parameters. Assessing Figure 4. The results shown in Figure 4 illumination of radio observatories and
impacts on astronomy also requires knowl- illustrate that the non-darkened Starlink radio-quiet zones does not interfere, indi-
edge of satellites’ apparent velocity, posi- satellites are brighter, and that the effec- vidually or in aggregate. Most importantly,
tion in the sky, and frequency of sightings. tiveness of the darkening treatment the work aimed to create a norm of coop-
decreases with increasing wavelength. eration within the astronomy community
Observations of the brightness magnitude and consideration of impacts on science
of the Starlink and OneWeb satellites have For completeness, observations of the and on the dark sky at the design and
been made for the purposes of a) deter- Starlink VisorSat (IAU 2020) measure a early regulatory approval of projects. The
mining how bright they are in various magnitude in the visible similar to that paper, which was co-signed by the IAU
astronomical spectral bands of interest, reported here for DarkSat. Observations and five countries (Chile, Ethiopia, Jordan,
and b) in the case of the Starlink satellites, of OneWeb satellites (Tregloan-Reed et Slovakia and Spain), was submitted to
determining how effective are the engi- al., in preparation), operating at an orbital the 58th session of the Science and Tech-
neering design changes implemented to height of 1200 kilometres, show that nical Sub-Committee of COPUOS held
decrease their brightness and meet the their brightness in the visible is about on 19–30 April 2021. Many COPUOS
recommendations of the SATCON1 (Walker 8 magnitude. However, in 46% of these members voiced support for the astron-
et al., 2020) and the Dark and Quiet Skies observations the satellites are brighter omy community’s concern and the need
for Science and Society (IAU 2020) work- than the SATCON1 recommendation of to find solutions, and the Committee
shops. SpaceX launched a mitigation 7.85 magnitude for a satellite at this agreed that the IAU should continue to
satellite, Starlink-1130a (dubbed DarkSat), altitude. study the matter further and report back
that included a special darkening treat- to the 2022 session.
ment of the Earth-facing sides of the
satellite structure. This was intended to

3 Figure 4. Brightness magnitude of Starlink satellites
Observations with the Chakana telescope Observations with VISTA telescope, 1130 (DarkSat) and 1113. Data extracted from
at the Ckoirama Observatory of the VIRCAM instrument, at Tregloan-Reed et al., (2020, 2021). The insert shows
Brightness magnitude at zenith and range = 550 km
Universidad de Antofagasta, Chile ESO’s Paranal Observatory the computed reduction in brightness for Starlink-1130
3.5 (DarkSat), similarly to Figure 6 in Tregloan-Reed et al.
(2021).
Percentage brightness reduction

4.5 55
50
5 45
40
5.5 35
30
6 25
Starlink 1113 500 1000 1500 2000
Starlink 1130 (DarkSat) Filter wavelength and passband (nm)
6.5
400 600 800 1000 1200 1400 1600 1800 2000 2200
Wavelength (nm)
Whilst COPUOS is an important forum in astronomy are one concern amongst Links
which to raise awareness and possibly many in this new megaconstellation era 1
ESO public statement: https://www.eso.org/public/
agree non-binding guidelines, a binding (Boley & Byers, 2021). This large number announcements/ann19029/
international treaty to implement the IAU of satellites in LEO creates a major con- 2
SpaceX’s VisorSat: https://spacenews.com/
policy guidelines is highly unlikely. Instead, cern in terms of orbital crowding, collision spacex-to-test-starlink-sun-visor-to-reduce-
the astronomy community must look to avoidance and control of debris. Whilst brightness/
3
national regulatory authorities. In this various space agencies are considering Report presented to the ESO Council: http://www.
eso.org/public/about-eso/committees/cou/cou-
regard, the primary issue is that, apart the problem, the existing space govern- 155th/external/Cou_1928_Satellite_
from a narrowly-focused law in the US ance system is being stretched to its lim- Constellations_161120.pdf
preventing “space billboards5” no states its. We hope the astronomy community, 4
Policy recommendations submitted to COPUOS:
have yet regulated optical pollution from along with all space actors and benefi- https://www.iau.org/static/publications/uncopuos-
stsc-crp-8jan2021.pdf
space, and the ITU does not include opti- ciaries of pristine dark and quiet skies, 5
US law on space advertising: https://www.law.
cal frequencies in its Radio Regulations. will work towards a shared stewardship cornell.edu/uscode/text/51/50911
A concern for radio astronomers is that of the night sky in a way that supports 6
Announcement of SATCON2 workshop: https://
so-called Radio-Quiet Zones, while offer- conservation, economic development, aas.org/posts/news/2021/05/satellite-
ing protection from ground-based radio science and exploration and sustainability constellations-2-workshop-announced
emitters, do not cover space-based radi- of the environment.
ation sources. Several activities to look Notes
for regulatory solutions are proceeding,
starting with the US-led SATCON2 pro- References a
Starlink 1130 satellite, NORAD ID 44932, COSPAR
ID 2020-001U
ject6 to explore the possibilities for legal Bortle, J. E. 2001, Sky & Telescope, February 2001 b
Starlink 1113 satellite, NORAD ID 44926, COSPAR
protection under environmental law, in Boley, A. C. & Byers, M. 2021, Sci Rep, 11, 10642 ID 2020-001N
addition to addressing national launch Char, F. et al. 2016, BAAA, 58, 200
and communications regulators and Curzi, G., Modenini, D. & Tortora, P. 2020, Aerospace,
7(9), 133
standards bodies. Dalton, G. B. et al. 2006, Proc. SPIE, 6269, 62690X
Emerson, J., McPherson, A. & Sutherland, W. 2006,
Despite the partial success of recent The Messenger, 126, 41
operator mitigations and some possibili- Hainaut, O. R. & Williams, A. P. 2020, A&A, 636,
A121
ties for introducing regulations to compel IAU 2020, Dark and Quiet Skies for Science and
operators to coordinate with astronomers Society — Report and Recommendations
and make design changes to satellites, Tregloan-Reed, J. et al. 2020, A&A, 637, L1
no combination of mitigations can fully Tregloan-Reed, J. et al. 2021, A&A, 647, A54
Walker, C. et al. 2020, Bull. AAS, 52(2)
avoid impacts on astronomy. With the Wood, L. 2003, Satellite constellation networks,
possibility of 100 000 satellites launching Internetworking and Computing over Satellite
in the coming decades, the impacts on Networks, (Boston: Springer), 13

The Organisation DOI: 10.18727/0722-6691/5238
Report on the Scientific Prioritisation Community Poll (2020)
Antoine Mérand 1 4
Department of Astrophysics, U niversity tered users targeted, 10% had accounts
Paola Andreani 1 of Vienna, Austria in both the ESO and European ALMA
Michele Cirasuolo 1 5
Hamburg Observatory, Germany portals, another 14% were registered in
Fernando Comerón 1 6
Institute for Astronomy, University the ALMA portal only, and the remaining
Itziar De Gregorio Monsalvo 1 of Edinburgh, Royal Observatory 76% were registered in the ESO portal
Miroslava Dessauges-Zavadsky 2 Edinburgh, UK only. Some 3700 email addresses,
Éric Emsellem 1, 3 7
INAF – Astrophysics and Space predominantly associated with the ESO
Rob Ivison 1 Science Observatory of Bologna, Italy portal, were invalid. From the remaining
Francisca Kemper 1 8
Institute of Astrophysics, KU Leuven, approximately 14 000 user accounts,
Franz Kerschbaum 4 Belgium 1673 complete responses were received,
Bruno Leibundgut 1 a response rate comparable to that of
Jochen Liske 5 the 2015 poll. The present poll was split
Ross McLure 6 ESO regularly updates its science- into three parts: 1) profile of respondent;
Tony Mroczkowski 1 driven perspective in order to provide 2) current and future observing facili-
Livia Origlia 7 the best facilities and services for its ties; 3) ESO in the coming decade. Here
Neil Philips 1 community. As part of this exercise, ESO we summarise the results and provide
Hugues Sana 8 polled its users between January and some highlights from the poll.
February 2020. Questions were inspired
by the previous poll, conducted in 2015,
1
ESO to probe any evolution of community Respondents’ profiles
2
Department of Astronomy, University of opinions and profile, with an emphasis
Geneva, Switzerland on the future of the Very Large Tele- The poll started with questions designed
3
CRAL, University Claude Bernard scope (VLT) and the VLT Interferometer to assess the profile of each respondent,
Lyon 1, ENS of Lyon, France (VLTI). Of the approximately 17 700 regis- including their academic and professional
What best describes your current position? If you have a PhD, for how many years have My home institute is best described as:
you had it (choose the closest number of years)?
5 years 2 years or less

Faculty member or 15.3% 9.4% Research institute
researcher (tenure- 29.1%
track or tenured Not applicable
10 years 15.6%
position)
16.3% Junior Observatory
51.7%
17.8%
Student 40 years
Graduate Mature or more
student Senior 3.8% Laboratory
12.6% 2.6%
15 years 30 years
Other
11.7% 11.6%
Researcher not in a 1.4%
tenure-track position Other 20 years University
31.0% 4.7% 16.3% 49.1%
My home institution is located in: Details for “Other”: Figure 1. (Above)

Respondents’ academic
Germany France Countries from which fewer than
profiles. Position (left),
13.7% 10.9% 10 responses were received
years post-PhD (centre)
Ireland Finland 5.3% and institution (right).
0.7% 1.0%
Italy ESO Brazil
9.9% 2.9% 2.6%
The Netherlands Denmark China
3.2% 1.0% 1.2%
Poland Czech Republic India
1.9% 0.7% 1.1%
Portugal Chile
1.0% 4.9% Japan
Spain Belgium 1.2%
5.9% 2.2%
Sweden Austria Mexico
2.0% 1.0% 1.0%
Switzerland Australia
2.9% 4.1% Russian
United Kingdom Other Federation USA Figure 2. Respondents’
10.3% 19.8% 1.0% 6.4% country of home institute.

background. As can be seen in Figure 1, My research activity focuses on:
the distribution shows a slight majority
of tenured or tenure-track researchers Other 2.5%
3.8%
1.7%
Senior
Mature
(51.7%), the rest consisting of non-tenured 1.0%
2.3% Junior
researchers (31.0%) and graduate stu- Simulation, modelling Student
29.8%
dents (12.6%). Regarding the number of or computational 31.1% 30.6%
years post-PhD (relevant for 84.4% of astrophysics 31.4%
31.6%
respondents), we have a relatively even 12.6%
spread between 2, 5, 10, 15, 20, 30 and Theory 12.2% 10.7%
11.8%
40 years. The respondents are predomi- 14.4%
nantly from universities (49%), with the 34.9%
remainder from research institutes (29%), Instrumentation 23.3% 24.1%
13.0%
observatories (18%) and laboratories (3%). 10.7%
90.0%
Observations 90.2% 91.5%
For later analysis, we define career 91.8%
stages (i.e., seniority) as: 2015 87.9%
– Respondent without PhD degree 0 20 40 60 80 100 0 20 40 60 80 100
(15.6%), of which 4/5 were students. Fraction of answers (%) Fraction of answers (%)
– Junior: up to 5 years post-PhD (24.7%)
(an average density of 4.9% per year I use data from these wavelength regimes:
post-PhD).
0.8%
– Mature: 5 to 20 years post-PhD (28.0%) Other 0.7% 0.6%
Senior
0.2% Mature
(an average density of 1.9% per year 0.5% Junior
post-PhD). Other messengers 5.3% Student
4.3%
– Senior: more than 20 years post-PhD (e.g. neutrinos, 4.5%
3.9%
cosmic rays, GW, etc.)
(31.7%) (an average density of 1.5% per 4.2%
34.7%
year post-PhD). 28.5% 28.8%
Radio 24.9%
21.9%
Our choices of the ranges of years after 40.8%
37.0%
a PhD do not have equal spread: if we mm/sub-mm 35.3%
32.4%
consider the average density per year post- 27.9%
28.5%
PhD, we seem to have obtained better Far-IR (30 to 23.7%
23.6%
response rates from junior people. 250 microns) 21.7%
17.2%
43.6%
Mid-IR (3 to
The poll collected information about home 30 microns)
36.0% 39.5%
30.0%
institutions. Only ESO Member States, 24.7%
78.3%
Australia as a Strategic Partner, the Host Near-IR (1 to 75.6%
3 microns) 71.4%
State Chile, and ESO were listed indi 67.9%
53.0%
vidually. Other countries could be entered 83.4%
manually. Figure 2 shows the relative Optical (0.3 to 84.0%
1 micron) 82.4% 84.3%
numbers of answers per country. More 73.5%
than 80% of the respondents are from Ultraviolet/

33.2%
34.4%
31.0%
ESO Member States, associated countries near-UV 30.2%
18.1%
and ESO. 25.1%
X-rays 20.6% 20.5%
18.4%
A multiple-choice question probed 15.3%
research categories: Observations; Instru- 9.2%
5.1%
Gamma rays 6.5%
mentation; Theory; Simulations; Other (Fig- 4.8%
4.7%
ure 3). The vast majority of respondents
0 20 40 60 80 100 0 20 40 60 80 100
(90.2%) are involved in observational Fraction of answers (%) Fraction of answers (%)
astronomy, with 10–30% pursuing other
types of research. Interestingly, a break- et al., 2015) is difficult, since then only a Figure 3. (Upper) Type(s) of research activities (left),
with a breakdown by career stage (right). In the left
down by seniority reveals that the Instru- single answer was allowed to this question.
panel, numbers for the 2015 poll are shown as the
mentation category is more strongly repre- However, the respondents then were also thinner and lighter grey bar. Note that, in 2015, only
sented amongst senior respondents (38%) dominated by observational astronomers. one answer was possible.
than amongst students (10.7%): compar-
atively fewer students participate actively Regarding the part of the electromagnetic Figure 4. (Lower) Wavelength domain(s) of
research activities (left) and broken down by career
in instrumentation development. Precise spectrum used (Figure 4), the majority
stage (right).
comparison with the previous poll (Primas of the respondents primarily focus on the

The Organisation Mérand, A. et al., Report on the Scientific Prioritisation Community Poll (2020)
optical and near-infrared, with a significant tant one decade from now by only 24.7% aimed at gauging what kind of ground-
number of respondents working in the and 16.9% of the respondents, respec- based capabilities respondents will need
submillimetre and radio domains. Only a tively. This difference between current in the 2030+ timeframe, with a maximum
small fraction of respondents use domains and future importance was also observed of three possible choices. The question
not covered by ESO telescopes (gamma five years ago, but only for stellar physics: was split between observing technique,
rays, X-rays, far-infrared, radio and other five years ago the number of people who spectral resolution and spectral domain.
messengers). A breakdown by career stage worked in Galaxy Evolution was equal to The results are presented in Figure 6.
shows that the more senior researchers the number who thought it would be an Polarimetry and high-contrast imaging
tend to use more multi-wavelength/multi- important field in the future. The research are the least selected (13% and 16%),
messenger facilities: students, junior, domains which have the largest difference whereas integral-field, multi-object and
mature and senior scientists use, on aver- between the number of respondents single-object spectroscopy and high-
age, 2.6, 3.2, 3.5 and 3.8 different spec- engaged in research in that domain and angular-resolution imaging (40%, 39%,
tral ranges and non-electromagnetic the number who think it will be very impor- 37% and 35%) are the most popular
messengers, respectively. tant in future are “Search for life outside techniques. Interferometry and wide-field
Earth”, “Planetary system formation and and/or low-angular-resolution imaging
Finally, we asked respondents in which evolution” and “Cosmology and/or funda- have intermediate results (26% and 25%).
domain they currently work, and how mental physics”. This is very similar to the
important they feel different fields will be 2015 results, again except for “Structure Comparing with the facilities offered by
in the 2030s (Figure 5). One can compare and evolution of galaxies (including AGNs)”. ESO (present and planned), we can iden-
how popular a field is currently with its tify the following missing capabilities:
perceived importance in the coming – High-resolution (R ~ 100 000 k) spec-
decade: the most popular fields, “Struc- Present and future facilities tropolarimetry in the visible: this is
ture and evolution of galaxies (including a capability offered outside of ESO,
AGNs)” (43.8%) and “Life cycle of stars” The second part of the poll focused on though not on 8-metre telescopes.
(40.6%), are predicted to be very impor- observing facilities. The first question – High-resolution (R > 10 000 k) interfer-
ometry in the visible and near-infrared;
What are your main areas of astrophysical research? high resolution (R > 10 000 k), high-
How important do you think these domains will be in the 2030s? contrast imaging in the near-infrared.
These two capabilities do not yet exist
Time domain
astronomy or and would offer a unique parameter
transients space.
The Sun and the
Solar System The second question concerned which
current ESO facilities are required for
Extreme states of future research (Figure 7). The respondents,
matter
of which 24% are registered at the ALMA
Search for life portal, and 86% at the ESO portal, indi-
outside Earth cated that they would most likely require
the VLT and the Extremely Large Tele-
Planetary system
formation and scope for their research (81.4% and 71.9%
evolution respectively), followed by the data archive
Life cycle of and the Atacama Large Millimeter/submilli
stars meter Array (ALMA) (58.9% and 49.5%,
Life cycle of
respectively). In 2015, the data archive was
interstellar not a possible choice. All facilities grew in
matter community interest. Amongst the facilities
Milky Way dynamics which grew in perceived importance, the
and evolution Atacama Pathfinder EXperiment (APEX)
Structure and jumped by a large factor; ALMA and the
evolution of galaxies VLTI roughly doubled in fractional answers.
(including AGN)
The large fraction of respondents (49.5%)
Large scale indicating that they intend to use ALMA in
structure of the
Universe
Cosmology and/or Figure 5. Current field of research (grey bars), and
fundamental 2020 Poll fraction of respondents thinking it will be very impor-
physics 2015 Poll tant in the future (green bar for increase and red bar
for decrease). The thinner and lighter colour bars
0 10 20 30 40 50 60 show the results of the 2015 poll. Note that in 2015,
Work currently in the field (%) it was not asked whether “Time domain…” will be of
(will be very important in the 2030s) importance in the coming decade.

Which of the following ground-based capabilities will be most important for your research in the 2030+ timeframe?
Polarimetry Integral-field spectroscopy
100 k or
more
Polarimetry 10 k to
(13%) 100 k
1 k to
10 k
1 k or
Integral-field less
spectroscopy NA
(40%)
Multi-object spectroscopy Single-object spectroscopy
100 k or
Multi-object more
spectroscopy 10 k to
(39%) 100 k
1 k to
10 k
1 k or
Single-object less
spectroscopy NA
(37%)
Interferometry High-contrast imaging
100 k or
Interferometry more
(26%) 10 k to
100 k
1 k to
10 k
High-contrast 1 k or
imaging less
(16%) NA
Wide-field and/or low-angular-resolution

High-angular- High-angular-resolution imaging (ู 0.5ೀ) imaging
resolution 100 k or
imaging (ู 0.5ೀ) more
(34%) Radio 10 k to
Sub-mm 100 k
3–30 μm 1 k to
Wide-field and/or 10 k
low-angular- 1–3 μm 1 k or
resolution imaging 0.4–1 μm less
(25%) 0.3–0.4 μm NA
0 5 10 15 20 0.3 to 0.4 to 1 to 3 to Sub-mm Radio 0.3 to 0.4 to 1 to 3 to Sub-mm Radio

Fraction of respondents (%) 0.4 μm 1 μm 3 μm 30 μm 0.4 μm 1 μm 3 μm 30 μm
Figure 6. On the left, the fraction of respondents or a telescope dedicated to transient fol- shop held in 2019 (Mérand & Leibundgut,
choosing each technique is shown (in percent, broken
low up. The other options each attracted 2019), the purpose of which was to dis-
down into wavelength domains). The right panels
show the density maps of the answers, in the “spectral about 25% of the respondents. All num- cuss future developments for the VLT and
resolution-spectral domain” plane (darker means more bers have at least doubled since 2015, VLTI, in the 2030+ timeframe. Several
answers). Areas covered by VLT and ELT instruments though only a single answer was allowed projects were discussed, and four pro-
are shown in yellow and purple respectively. Green
in the previous poll. jects were selected by the Science and
areas cover ALMA (only for p olarimetry and interfer-
ometry, even though ALMA covers other aspects). Technical Committee (STC) for further
We also polled the community regard- review: BlueMUSE, GRAVITY+, HR-MOS,
the future is remarkable, given that only ing planned facilities, ground- or space- and SPHERE+ (in alphabetical order).
24% of the users polled are currently based (Figure 7, right panel). The answers Respondents were asked how relevant
registered at the European ALMA portal. are very similar to those in 2015, with the those projects were for their research.
Only 4-metre-class telescopes seem to James Webb Space Telescope (JWST), The breakdown of answers by instrument
have remained stable since 2015. There Vera C. Rubin Observatory (formerly the is given in Figure 8.
do not seem to be any strong genera- LSST), Nancy Grace Roman Space Tele-
tional trends in these answers. scope (formerly WFIRST) and the Square Something the Figure does not show is
Kilometre Array (SKA) (in that order) that 67% (26%) of the respondents find at
Regarding possible future ground-based included as a choice in the 2015 poll. least one (two) of the four projects “very
facilities (Figure 7, lower left), a large relevant”. This percentage climbs to 89%
majority (75%) expressed the need for a (69%) when “relevant” is also included.
dedicated spectroscopic telescope. It ESO in the coming decade This means that at least one VLT2030
should be noted that the poll did not instrument captured the interest of the
define what “dedicated” meant: it could We began by asking a question that vast majority of respondents. The next
either be a highly multiplexed telescope, referred back to the “VLT in 2030” work- question attempted to identify capabilities

Which ESO facilities do your future research objectives require? Which other planned facilities are essential for your future research?
Other 2.2% Other 8.0%
Data Archive 58.6% SPICA (Space Infrared

Telescope for Cosmology 26.8%
ELT (40-m-class and Astrophysics)
optical/IR telescope) 71.9%
ngVLA (next 21.3%
VLTI (1.8-m or 8-m generation VLA)
telescopes 22.1% ARIEL (Atmospheric
interferometer) Remote-sensing 10.2%
Infrared Exoplanet
VLT (8-m-class 81.4% Large survey)
optical/IR telescopes)
MSE (Maunakea
< 4-m optical/IR Spectroscopic 21.6%
38.1% Explorer)
telescopes
LISA (Laser
APEX (single-dish 14.1% Interferometer Space 8.7%
sub-mm) Antenna)
WFIRST (Wide Field
ACA (mm/sub-mm 17.1% Infrared Survey 35.7%
compact array)
Telescope)
Euclid
ALMA (mm/sub-mm (Spectroscopic All Sky 25.0%
array) 49.5%
2015 Cosmic Explorer)
ATHENA (Advanced
0 20 40 60 80 100 Telescope for High 16.6%
Fraction of answers (%) ENergy Astrophysics)
PLATO (PLAnetary
Which of these possible facilities do your future Transits and 18.1%
research objectives require? Oscillations)
CHEOPS
Other 6.9% (CHaracterising 12.0%
ExOPlanet Satellite)
Very long baselines 23.0% JWST (James Webb 70.4%
in sub-mm Space Telescope)
Large mm/sub-mm
single dish (30–50 m 22.9% CTA (Cherenkov 5.3%
in diameter) Telescope Array)
Optical/IR SKA (Square
interferometer with 26.0% 28.8%
Kilometre Array)
kilometric baselines
Dedicated optical/IR LSST (Large Synoptic
spectroscopic 10+-m 75.1% 44.4%
Survey Telescope)
telescope 2015 2015
0 20 40 60 80 100 0 20 40 60 80 100
Fraction of answers (%) Fraction of answers (%)
Figure 7. Importance of present and future (planned The next question concerned current tend to be more in favour of Visitor
or not) facilities. The thinner and lighter grey bars show
observing modes and scheduling capa- Mode than their mid-career peers.
the numbers from the 2015 survey (where available).
bilities and which of those will be impor- – V isitor Mode and Delegated Visitor
tant for research objectives in the coming Mode a have exactly the same level of
missing from the current and planned decade. The question is similar to one support.
offerings at ESO’s VLT/I. Nearly a quarter asked in 2015, but the numbers cannot – Director’s Discretionary Time (DDT) is
of the respondents (399 or 24%) com- be compared directly as the methodology found to be less relevant to students
mented, most of them providing details was slightly different. However, we can than to more senior researchers.
of missing capabilities. Most answers make some interesting observations (see
suggest different telescopes (for example, Figure 9): Regarding possible future operational
smaller telescopes or 10-metre spectro- – Normal Programmes and Service Mode capabilities, most ideas attracted a posi-
scopic telescopes). In broad terms, the have the most support (~ 90% find it tive response (see Figure 10). The most
missing capabilities for the VLT are in the very relevant or relevant). Public Surveys favoured options are those bringing
areas of multi-object spectrographs and and Large Programmes also have sig- operations towards more virtual access:
integral-field units (99 answers), in the near- nificant support (~ 70%). Around half archival proposals, remote observations,
infrared, or concern adaptive optics at (40–60%) of the community found the cloud-based access to data and reduction
bluer wavelengths. Other popular requests other modes relevant. tools. Many new features have generational
are for VLTI extended capabilities (48 – V isitor Mode is less favoured (55.5%) trends. The ideas that are clearly preferred
answers), and polarimetric modes, either compared to service mode (88.9%). by younger researchers include distributed
for spectroscopy or imaging (34 answers). There is a generational trend: the peer review, dual-anonymous proposals,
youngest and most senior scientists the possibility of applying for several facil-

In June 2019 ESO organised a workshop to discuss future developments for VLT and VLTI in ities at once, and cloud-based access to
the 2030+ timeframe. Several projects were discussed and four projects were selected by data and pipelines. Only the capability for
the Science and Technical Committee (STC) for further review. Three of these projects have
advanced design, whereas the fourth is only a concept, currently lacking a consortium. remote observations is less favoured by
How relevant are the following projects for your research? younger scientists; this trend was also seen
in the level of enthusiasm for the Visitor
Development of a High-resolution (R ~ 50 000 k)
Multi-Objects Spectrograph
29.2% 26.5% 24.4% 19.9% Mode amongst more junior researchers.
SPHERE+: improved extreme AO and 24.8% 29.6% 21.5% 24.0%
spectroscopic capabilities
GRAVITY+: VLTI with improved Cross analysis: facilities versus
18.2% 21.2% 26.0% 34.6%
sensitivity and dynamic range research fields
BlueMUSE: integral-field 27.8% 24.6% 24.8% 22.8%
spectroscopy in the blue Although the poll was fully anonymous
0 20 40 60 80 100 we can match answers to one question
Very relevant Relevant Mildly relevant Irrelevant with those to another, as we already did
to analyse answers by career stage. For
Which of the current observing modes and scheduling capabilities will be important for example, we can examine the desire for
your research objectives in the coming decade? current and future facilities as a function of
Director’s Discretionary Time (immediate; hot
the research field the respondent works in.
33.2% 32.7% 20.7% 13.4%
topic; quick follow up; feasibility of risky obs.)
Simultaneously in time with other facilities 19.2% 23.4% 21.5% 35.9% Figure 11 displays both coloured and tab-
(e.g. XMM)
ulated values for the people finding cur-
Delegated Visitor Mode (observations at a
fixed date, with remote connection)
23.3% 32.1% 26.8% 17.8% rent, planned or future projects “relevant”
Visitor Mode (observations at the telescope, or “very relevant” for their research, bro-
23.1% 32.4% 29.1% 15.4%
at a fixed date) ken down by community. The data are
Service Mode 57.6% 31.1% 8.0% very rich, but a few broad conclusions
(condition adapting queue observing)
can be drawn: the VLT, the ELT, the data
Public Surveys 41.4% 30.2% 19.8% 8.5% archive and a future 10-metre spectro-
Large Programmes (more than 100 hours, scopic telescope are considered to be
45.9% 32.1% 16.4% 5.6%
over several semesters) true multi-purpose machines, since they
Normal Programmes (less than 100 hours) 60.8% 33.0% 5.2% are embraced by a large fraction of all
communities. Other facilities are special-
Monitoring Programmes (limited time over
several semesters)
24.9% 32.4% 17.7% 25.0% ised in terms of the communities they
serve, because of their excellence in spec-
Rapid response and target of opportunity 22.4% 20.3% 21.8% 35.6%
tral coverage, angular resolution, sensitivity,
0 20 40 60 80 100 operational modes, etc. The most popu-
Very relevant Relevant Mildly relevant Irrelevant lar research area is “Structure and Evolu-
tion of Galaxies (including AGN)” with 733
How favourably do you consider the following possible capabilities? respondents indicating that they work in
this area. The facilities that these respond-
Archival proposals (i.e. to obtain technical support
to perform re-analysis of data in the archive)
77.1% 19.7% 3.2% ents would like to use for their future
Cloud-based access to raw data research closely match the overall outcome
and reduction pipelines (instead of downloading 71.2% 23.8% 5.0%
data and installing pipelines on your computer) of the poll. The second-most popular
Condition-adapting Visitor Mode (visitor in research area is “Life Cycle of Stars”, where
standby or remote, accessing the telescope 52.7% 39.0% 8.3%
the optical telescopes (VLT and ELT) and
only under the very best observing conditions)
Remote observations (Visitor Mode without the archive are mentioned as important
71.6% 23.1% 5.3%
travelling to observatories but with full interactivity) facilities for future research, while the third-
Possibility to apply for time jointly to ALMA 51.3% 42.4% 6.3% most p opular research area “Life Cycle
and JWST/HST on a single proposal
of Interstellar Matter” clearly benefits from
Possibility to apply for time jointly to VLT
and ELT on a single proposal
61.3% 34.8% 3.9% having access to ALMA, which for these
Possibility to apply for time jointly to ELT and 46.8% 46.7% 6.5%
respondents is as important as the VLT
other ELTs on a single proposal to achieve their research goals.
Possibility to apply for time jointly to VLT 52.3% 43.0% 4.7%
and ALMA on a single proposal
Dual anonymous review of observing proposals 62.1% 25.5% 12.4%
Figure 8. (Upper) Relevance of the VLT 2030+ instru-
Distributed peer review of observing proposals 58.0% 32.9% 9.2%
ments according to the poll respondents.
Fast turnaround observing programmes Figure 9. (Middle) The relevance to respondents’

45.6% 36.5% 17.9% research of current observing modes and scheduling
(call every few months)
capabilities.
20 40 60 80 100
In favour Neutral Against Figure 10. (Lower) Possible future capabilities.

What are your main areas of astrophysical research? How relevant are the following Which of these possible
projects for your research? facilities do your future
ൻ Which ESO facilities do your future research objectives require? (very relevant or relevant) research objectives require?
Cosmology and/or
fundamental physics 54% 15% 16% 39% 84% 18% 76% 53% 10% 66% 38% 41% 56% 78% 20% 28% 22%
N = 312 (19%) 167 48 50 121 262 57 238 164 32 206 118 129 174 244 63 88 68
Large-scale structure
of the Universe 65% 21% 19% 34% 84% 14% 77% 55% 10% 77% 29% 35% 55% 80% 16% 35% 19%
N = 294 (18%) 192 61 56 99 246 40 225 161 29 225 84 102 162 236 47 103 56
Structure and evolution of

galaxies (including AGN) 63% 20% 16% 31% 83% 16% 75% 58% 10% 72% 32% 41% 56% 79% 21% 30% 26%
N = 733 (44%) 463 144 120 230 611 115 547 428 71 525 232 297 408 578 156 222 192
Milky Way dynamics

and evolution
N = 388 (23%)
33% 14% 12% 45% 83% 16% 73% 64% 7% 58% 37% 49% 73% 82% 21% 18% 18%
127 54 48 176 322 62 284 250 28 224 143 192 282 318 81 69 69
Life cycle of
interstellar matter
N = 462 (28%)
72% 35% 29% 31% 72% 21% 65% 58% 15% 55% 38% 50% 51% 63% 26% 40% 36%
333 160 134 142 331 98 299 269 68 256 176 232 235 290 122 187 168
Life cycle of stars

N = 679 (41%) 40% 15% 14% 46% 81% 28% 71% 61% 9% 47% 44% 61% 63% 73% 30% 17% 21%
274 104 94 315 552 188 479 416 60 319 301 412 429 499 207 113 140
Planetary system
formation and evolution
N = 476 (28%)
45% 17% 11% 42% 82% 36% 75% 56% 7% 28% 54% 79% 50% 68% 36% 17% 25%
216 81 50 198 389 169 357 265 31 131 259 375 236 326 171 79 121
Search for life

outside Earth
N = 252 (15%)
37% 10% 9% 41% 83% 35% 79% 57% 4% 25% 55% 85% 54% 72% 38% 15% 19%
92 26 22 103 209 88 200 143 11 64 139 214 136 182 96 37 49
Extreme states
of matter
N = 110 (7%)
44% 11% 14% 52% 81% 23% 66% 55% 16% 50% 54% 51% 45% 70% 23% 17% 25%
48 12 15 57 89 25 73 61 18 55 59 56 49 77 25 19 27
The Sun and the

Solar System
N = 179 (11%)
44% 17% 14% 50% 77% 25% 65% 68% 12% 35% % 70% 54% 69% 31% 21% 26%
78 30 25 89 137 44 117 122 22 62 78 126 97 124 55 38 46
Time domain astronomy

or transient
N = 470 (28%)
40% 10% 9% 56% 87% 25% 74% 64% 10% 50% 44% 60% 56% 80% 26% 15% 19%
188 45 40 263 409 118 347 303 49 236 208 281 265 378 124 69 90
in sub-mm
BlueMUSE
Large mm/sub-mm single-

HR-MOS
N = 1257 (75%)
Optical/IR interferometer
with kilometric baselines
N = 435 (26%)
None relevant
N = 176 (11%)
N = 877 (52%)
GRAVITY+
N = 658 (39%)
SPHERE+
N = 911 (54%)
N = 931 (56%)
Dedicated optical/IR
spectroscopic 10+-m telescope
dish (30–50 m in diameter)

N = 383 (23%)
Very long baselines
N = 384 (23%)
ALMA (mm/sub-mm array)
N = 828 (49%)
ACA (mm/sub-mm compact

array) N = 286 (17%)
APEX (single-dish sub-mm)

N = 235 (14%)
VLTI (1.8-m or 8-m telescope

interferometer) N = 370 (22%)
ELT (40-m-class optical/IR

telescope) N = 1203 (72%)
N = 637 (38%)
< 4-m optical/IR telescopes
VLT (8-m-class optical/IR

telescopes) N = 1362 (81%)
Data Archive
N = 981 (59%)
Figure 11. Interest in facilities, broken down by astro- latter. The final 10% of users in the pool this poll but did not feature in the 2015
physical fields. Facilities are in columns (current facili-
have accounts in both portals. The version), as well as for data-reduction
ties to the left, VLT2030+ projects in the middle, future
projects to the right). Percentages are computed by respondents indicate a high demand for support (Figure 10). Concerning future
rows (research fields): for example, 72% (N = 333) of the VLT, ALMA and the data archive, as facilities, the broadest interest is in a ded-
people working in “Life cycle of the interstellar matter” well as the ELT and future operational icated spectroscopic telescope. In con-
(N = 462) find ALMA relevant. Colours reflect the per-
modes. This is testimony to the relevance clusion, the working group felt that the
centages, while coloured areas are proportional to
the number of answers. For example, among all com- of these facilities, and an indication that polled community values and demands
munities, “Life cycle of the interstellar matter” has the users are engaged in their long-term use. the broad diversity of tools that ESO
largest fractions interested in ALMA (333/462 = 72%), Most of the observational parameter operates and strongly supports existing
even though “Structure and evolution of Galaxies, inc.
space of future interest is served by exist- and planned ESO facilities.
AGNs” attracts more people (463), but a smaller frac-
tion of the community (463/733 = 63%). ing and planned ESO facilities (Figure 6).
Yet two useful windows in this parameter
space not currently covered by ESO are References
Conclusion apparent: high-resolution spectropolarim- Mérand, A. & Leibundgut, B. 2019, The Messenger,
etry and high-resolution, high-angular- 177, 67
The 2020 ESO community poll reached resolution (interferometry and high- Primas, F. et al. 2015, The Messenger, 161, 6
a diverse community spread across career contrast imaging) in the optical and near-
stage, research field and wavelength infrared. There is broad interest in the Notes
regimes, revealing the broad use of and three new instruments proposed at the
interest in ESO facilities. The respond- VLT2030 workshop, as well as in a a
Designated Visitor Mode means that the
ents come from a pool of ESO and Euro- high-resolution multi-object spectro- astronomer connects remotely to the observatory,
rather than visiting in person.
pean ALMA portal users, with 76% com- graph. The poll shows a strong demand
ing from the former and 14% from the for the data archive (which was added in

Instrumentation
This lone antenna is part of the Atacama Large

Millimeter/submillimeter Array (ALMA), a telescope
that comprises 66 high-precision antennas spread
out across the Chajnantor plateau, located high up
in the Chilean Andes. In this image the spectacu-
larly multi-coloured view of the sky above ALMA is
ESO/P. Horálek
in full display: green airglow hovers above the hori-

zon, the Large Magellanic Cloud peeks out from
behind the antenna, and the magnificent sprawl of
the Milky Way stretches out overhead.
Instrumentation DOI: 10.18727/0722-6691/5239
A Guide to ALMA Operations and Interactions with

the Community
Martin Zwaan 1 partnershipa spread over four continents, Currently, there are four global IXTs:
Evanthia Hatziminaoglou 1 in which the European partner is ESO, Engineering, Computing, Science, and
Francisca Kemper 1 representing its Member States. The Science Operations. An IXT dedicated
Leonardo Testi 1 three executives that jointly operate ALMA to the coordination of the major ALMA
Elizabeth Humpreys 1, 2 are ESO, the AUI/National Radio Astron- system upgrades, the Integrated Devel-
Misato Fukagawa 3 omy Observatory (NRAO) and the National opment Team, is in the process of being
Anthony Remijan 4 Astronomical Observatory of Japan set up. Staff in the regions and in Chile
Andy Biggs 1 (NAOJ). On-site operations are executed are responsible for the full distributed
María Díaz Trigo 1 by a shared team based in Chile headed effort towards the goals and tasks within
Fabrizia Guglielmetti 1 by the Joint ALMA Observatory (JAO) each IXT.
Eelco van Kampen 1 Director and Deputy Director and jointly
Luke Maud 1 supported by the executives. In contrast, Here we focus in particular on two IXTs.
Anna Miotello 1 all off-site activities (related to science The first, the Integrated Science Opera-
Dirk Petry 1 operations, maintenance, and develop- tions Team, or ISOpT, is tasked with defin-
Gergö Popping 1 ment) are conducted regionally, with each ing and optimising the end-to-end work-
Suzanna Randall 1 executive providing services and support flow of science operations, from proposal
Thomas Stanke 1 tailored to the needs of its regional commu- preparation to data reduction and user
Felix Stoehr 1 nity and following the regional processes. support. The team develops and imple-
ments policies governing global science
As one of ESO’s primary facilities, off-site operations and is responsible for deliver-
1
ESO ALMA operations are embedded in ESO’s ing the full package of science operations
2
Joint ALMA Observatory, Santiago, organisational structure, within the ESO tasks. The ultimate goal is to provide an
Chile ALMA Support Centre, a Division of the effective, efficient and homogeneous
3
National Astronomical Observatory of ESO Directorate of Operations, working global user experience in order to maxim-
Japan, Tokyo, Japan closely with the ESO Directorates of Sci- ise high-quality and transformational sci-
4
National Radio Astronomy Observatory, ence and Engineering. The ALMA Support ence. The team also acts on issues that
Charlottesville, USA Centres in the three regions provide all are raised through various channels,
interactions between the regional commu- including ALMA staff from the ALMA
nities and ALMA and are responsible for Regional Centres (ARCs) and the JAO.
A primary goal of the Atacama Large the execution of the ensemble of regional
Millimeter/submillimeter Array (ALMA) activities related to ALMA operations. The management of each IXT within
has always been to be a facility accessi- These activities are related to science ALMA is composed of one member from
ble to astronomers, radio-interferometry operations, computing, engineering, the JAO and three from the regions. In
experts and non-experts alike. As a development, and public outreach. The the case of ISOpT, this means the head of
project, it is strongly committed to lis- heads of the three regional ALMA Sup- the ALMA Department of Science Opera-
tening to its users and to utilising this port Centres, together with the JAO Dep- tions (DSO) at the JAO and the managers
input in decision making and priority uty Director, form the ALMA Management of the three ARCs. In the decision making
setting. Feedback from the community Team (AMT). The AMT is the forum in process, ISOpT usually invites input from
often highlights the perceived complex- which any issues relating to the planning experts and aims to reach a unanimous
ity of ALMA’s organisational structure and management of ALMA operations decision amongst the members.
and, by extension, a diffuse uncertainty are shared, discussed, coordinated and
around how to make users’ voices resolved. The ultimate authority over Practices related to proposal preparation/
heard. The aim of this article is to pro- ALMA high-level principles and policies is submission (Phase 1), programme prepa-
vide insight into the functioning of ALMA the ALMA Board. As the overall ALMA ration/execution (Phase 2), project track-
as an integrated observatory, with an governing body, it is the primary forum ing, data reduction, data delivery and
emphasis on science and science oper- for interactions amongst, and decisions user support at each of these steps are
ations. We present information on the of, the three ALMA contracting parties. all explicitly under the remit of ISOpT. It
ways the observatory communicates also deals with innovative improvements
with the broader community, with a to observing strategies, increased pipe-
focus on the mechanisms by which the The integrated teams line reliability, data processing issues and
community can provide feedback to the improvements. An optimal information
project. One of the main challenges to the smooth flow to the science community regarding
functioning of ALMA is the coordination science operations is another priority, as
between the JAO and the regions such is supporting the user community through
ALMA organisational structure that effective support of operations, main- the Helpdesk, face-to-face consultations,
tenance, and development is ensured. training events and community days.
To set the scene, below we provide a To this end, for each major area an Inte- Staff across the three ARCs (and related
very brief overview of ALMA’s organisa- grated Team (IXT1) was created and ARC nodes) and the DSO jointly carry out
tional structure. ALMA is a worldwide tasked with executing this coordination. these tasks.

The Integrated Science Team (IST) is Figure 1. The rede-
signed ALMA Science
formed of the three regional programme
Archive query interface
scientists and the observatory scientist. provides access to all
Responsibilities of the IST include advising p roprietary and public
ALMA on the scientific priorities for the ALMA data.
ALMA Development Program and provid-
ing scientific support for this programme.
The IST also monitors the scientific pro-
ductivity of ALMA and proposes ways to
improve it. It also organises the triennial
international ALMA science conference,
and supports the organisation of topical
or regional scientific and development
meetings.
ALMA Subsystems
A primary responsibility of ISOpT is to set munity and welcome user feedback as Rendering Tool for Astronomy (CARTA)
requirements for science software sub- an essential facet of setting requirements. remote visualisation and the new ALMA
systems used by the scientific community The subsystem scientist and their working virtual observatory (VO) services. Finally,
and ALMA staff, and to verify that this group weigh the relative importance of all the ALMA Observing Tool is being
software meets the design specifications new feature requests and improvements upgraded to a web-based application.
and scientific requirements. To this end, and discuss solutions with the computing The current desktop tool began develop-
ISOpT interacts closely with the Inte- team, who are then tasked with their ment nearly twenty years ago and so
grated Computing Team (ICT) in the area timely and robust implementation. the upgrade will also bring it up to date in
of software development priorities, test- terms of the technologies used, introduc-
ing, and deployment. ISOpT also over- Three of the subsystems listed as examples ing both user enhancements and easier
sees the testing of these subsystems and above are led from the ESO ARC. SnooPI maintainability. It should be noted that the
preparation of end-user documentation. is now a very mature system and is one Observing Tool upgrade, upcoming
of the main interfaces between the user ALMA Science Archive enhancements,
For each software subsystem, a subsys- community and the observatory. Principal and the Additional Representative Images
tem scientist, who is an expert on that Investigators (PIs) and Co-Investigators for Legacy (ARI-L 2) project, which pro-
subsystem at one of the ARCs or at the (Co Is) can log into the tool to follow their duces homogeneous imaging products
JAO, leads the formulation of the scien- ALMA projects from the moment of for Cycles 2 to 4, are all possible thanks
tific requirements and priorities. It is of proposal submission all the way to data to the ALMA Development Program3.
course essential that these priorities are delivery, and they can download quality
aligned with the overall ALMA observa- assurance reports. The ALMA Science
tory priorities. To ensure that all ALMA Archive is developing very fast to meet Department of Science Operations
regions are represented in the collection user requirements. Some recent high-
and prioritisation of requirements, the lights include a completely redesigned Although ISOpT is overseeing the overall
subsystem scientists are supported by query interface, the Cube Analysis and global ALMA science operations, it is
working groups with members from all
four regions. In general, consensus must
be reached in the working groups before
improvements or new feature requests
are raised to the implementation phase.
Amongst the numerous subsystems,

some are relatively invisible to the user
community, such as internal tools to track
observing projects and quality assurance
or scheduling, and their requirements are
defined through well-established proce-
dures within the project. Others, such as Figure 2. SnooPI allows
the ALMA Observing Tool, the Snooping PIs, Co-Is and D elegees
to track the observa-
Project Interface (SnooPI), the Science tional and processing
Archive, the Science Portal and the Help- status of their ALMA sci-
desk, interact directly with the user com- ence projects.

Instrumentation Zwaan, M. et al., A Guide to ALMA Operations and Interactions with the Community
important to clarify that many of the tasks ALMA users. The face-to-face user sup- From the community to the project
within science operations are the explicit port, one of the backbones of ALMA’s
responsibility of the DSO at the JAO. This user support structure, is fully delegated ALMA is very much committed to nurtur-
department is directly responsible for to the nodes in the ARC network. Euro- ing an optimal two-way communication
the day-to-day science operations at the pean users can receive face-to-face or with its users. News from the observatory
telescope, including the execution of the virtual help from ALMA experts at each is posted on the ALMA Science Portal4,
science projects, but also troubleshoots of the seven nodes. The functioning and in Europe this information, as well as
issues with the ALMA array, acquires of this network is explained in detail by additional ALMA news relevant to the
and processes test data, performs trend Hatziminaoglou et al. (2015). European community, is disseminated
analysis of the array performance, and through a newsletter5 that is distributed
coordinates quality assurance. Issues One particular priority of the ESO ARC to all European ALMA users. Further-
related to antenna availability, configura- is to identify synergies with departments more, most of the European ARC nodes
tion details (pad availability, the relocation responsible for user support and science maintain their own regional newsletters.
schedule) are also within the purview of operations at ESO’s other facilities. For Everything our users need to be aware of,
the DSO. Staff within the DSO are respon- example, the ALMA Helpdesk has migrated for example information related to calls
sible for most of the astronomer-on-duty to the new service provider Deskpro. for observing proposals, data reduction,
shifts, augmented by visits from ARC This same tool is now also employed for community events, and developments in
astronomers. The yearly process that La Silla Paranal, which means that all capabilities, is shared through these
makes new observing capabilities possi- ESO users will have a uniform user expe- channels, as well as via the various social
ble — known as ObsMode — is also led rience when creating tickets and interact- media accounts.
and coordinated by the DSO (Takahashi ing with support staff.
et al., 2021). Equally important though, is that ISOpT
As another example, ALMA data products is fully aware of its users’ needs and
have now been integrated into the ESO wishes and is in touch with what the
ALMA Regional Centres Archive Science Portal along with data community considers important to reach
products from the ESO’s other facilities. their scientific goals (see Figure 3). The
Within the regions, ALMA science opera- Millions of datasets can be browsed jointly front-line channel for all feedback to
tions are organised at the ARCs. In addi- through a uniform set of query items, pro- ALMA is the Helpdesk, although in reality
tion to hosting a subset of the subsystem viding a unique integrated panchromatic most information that comes to ALMA
scientists, the ARCs are responsible for view of the southern hemisphere, extend- through this channel is directly related to
user support in their regions and work ing from the near-ultraviolet to millimetre observing programmes, for example sta-
closely together with their colleagues in wavelengths. The selected data can then tus of observations, requests for help
Chile and the other ARCs to deliver the be downloaded from the respective por- with data reduction. Another very impor-
science operations programme. tals for ALMA and ESO. tant channel is provided by advisory
committees, through which the commu-
The ARCs also host complete mirrored nity can make its voice heard on topics
copies of the ALMA Science Archive,
one additional copy of which is located
in Chile. It contains all raw science target ALMA science operations
and calibration data, all data products
produced by the ALMA Pipeline or by
manual data reduction, and quality assur-
ARC nodes
ance parameters. Externally contributed
ARC
products (for example, from Large Pro- ESO
grammes) will also be available in the
near future. The ARCs contribute to doc-
umentation, deliver call-for-proposal and
ALMA Board
observation preparation support, and
assist with enabling new capabilities (see
Maud et al., 2021). Data processing and ASAC ESO STC
quality assurance (see Petry et al., 2020)
constitute a major task as well, in addition ESO Users Figure 3. An illustration
of how community input
to Helpdesk staffing. Committee
can reach ALMA sci-
ESAC
ence operations. Euro-
In Europe, the ARC is uniquely organised pean ALMA users can
as a distributed structure, where, in addi- submit Helpdesk tickets,
ALMA Helpdesk
contact their ESAC or
tion to the ARC department at the ESO UC representatives, or
headquarters, seven nodes throughout European ALMA users communicate directly
Europe provide enhanced services to with the ARC nodes.

140 ing with its user community. This article
Cycle 8
proposal deadline has given an overview of ALMA’s organi-
Number of requested MOUSs
Shutdown sational structure, its functioning as an

105 integrated observatory and the different
channels available to the European ALMA
users to provide their feedback and make
70 themselves heard, helping to keep ALMA
at the forefront of astronomical research.
35
References
Cioni, M. et al. 2019, The Messenger, 176, 8

0 Hatziminaoglou, E. et al. 2015, The Messenger,
0 0 0 0 0 0 0 0 0 0 1 1 1 1 1
/2 2/2 3/2 4/2 5/2 6/2 7/20 8/2 9/2 0/2 1/20 2/2 1/2 2/2 3/2 4/2 5/2 162, 24
01 0 0 0 0 0 0 0 0 1 1 1 0 0 0 0 0 Maud, L. et al. 2021, The Messenger, 183, 13
Date Petry, D. et al. 2020, The Messenger, 181, 16
Takahashi, S. et al. 2021, ALMA Memo, 618
Figure 4. The number of member observing unit sets
(MOUSs) that the ESO ARC delivers through its cali-
brated data set service, as a function of time.
Links
1
The ALMA management structure: https://www.
related to operations, science priorities, broad range of scientific competences. almaobservatory.org/en/about-alma/the-people/
enhancements and future development. As an example of the recommendations alma-management/
that are formulated, following signals from 2
The ARI-L project: https://almascience.eso.org/
alma-data/aril
The European Science Advisory Commit- the community, the UC made a strong 3
The ALMA Development Program: https://www.
tee (ESAC), a subcommittee of the Scien- recommendation to ESO to implement a almaobservatory.org/wp-content/uploads/2018/07/
tific Technical Committee (STC), is tasked service that provides access to ALMA 20180712-alma-development-roadmap.pdf
with advising ESO on all technical and calibrated measurement sets for archive 4
The ALMA Science Portal: http://almascience.org
scientific matters related to ALMA. The users and PIs. The ESO ARC followed up 5
The European ARC newsletter: https://www.eso.
org/sci/facilities/alma/news/arc-newsletter.html
ESAC members, all of whom are repre- on this recommendation and implemented 6
ASAC members: https://www.almaobservatory.
sentatives of the European ALMA science this service, which is now a very popular org/en/about-alma/the-people/the-alma-
community, are appointed by ESO, and welcome addition to the suite of ser- committees/asac-members-list/
t ypically for a period of three years. The vices it offers (see Figure 4). 7
ESAC members: https://www.eso.org/public/
ESAC is coordinated by the European about-eso/committees/stc-esac/stc-esac2021.html
8
ESO Users Committee: https://www.eso.org/
ALMA Programme Scientist, who inter- ALMA is constantly seeking out new ways
public/about-eso/committees/uc/uc2021.html
faces between the committee and the of receiving feedback from, and engaging
ALMA activities at ESO. Counterparts of with, its user community. The latest such
the ESAC exist also in North America and effort is the ongoing Re-defining the User Notes
East Asia, and representatives of all these eXperience (RedUX), a series of interviews a
ALMA is a partnership of ESO (representing its
three regional committees have seats on with individual ALMA users that expressed Member States), the National Science Foundation
the ALMA Science Advisory Committee their interest in participating, following a in the USA and the National Institutes of Natural
(ASAC6), which, in turn, advises the ALMA call for volunteers in the autumn of 2020 Sciences in Japan, together with the National
Board. The ALMA observatory scientist sent to ALMA users world-wide. The aim Research Council (Canada), the Ministry of
Science and Technology and the Academia Sinica
coordinates the ASAC and interfaces of this exercise is to take a holistic view Institute of Astronomy and Astrophysics (Taiwan)
between the ASAC and the JAO. Each of of the ALMA user experience, identify and the Korea Astronomy and Space Science
the members of the ESAC7, as well as the issues and implement innovative solutions, Institute (Republic of Korea), in cooperation with
European ALMA Programme Scientist, can in an effort to further enable high-impact the Republic of Chile.
be contacted at any time in relation to science with ALMA.
ALMA development and science matters.
European ALMA users have yet another Concluding remarks

channel through which to have their say
in ALMA priorities and directions. The ALMA is a unique project in the astro-
ESO Users Committee (UC8; see also nomical landscape, a large collaboration
Cioni, 2019) is an advisory body that of diverse institutes and cultures, all
represents active European users of the bringing together the best of their worlds.
La Silla Paranal Observatory, including the In order to achieve one if its primary
Atacama Pathfinder EXperiment (APEX), goals, i.e. the accessibility of the facility
and ALMA. Its members are appointed to all astronomers, regardless of their
for a three-year term and encompass a expertise, ALMA is committed to engag-

Instrumentation DOI: 10.18727/0722-6691/5240
Upgrade Strategies for the ALMA Digital System
Benjamin Quertier 1 tion (see, for example, Kemper, 2020). baseband compared to 256 without the
Stéphane Gauffre 1 Bringing the signals from up to 66 anten- use of the TFBs).
Antsa Randriamantena 1 nas together in the correlators was one
Marina Studniarek 1 of the major technological challenges of Whilst ALMA inspired major progress in
Carlos De Breuck 2 ALMA. It involved laying more than 1000 (sub-)millimetre technology during its
Tony Mroczkowski 2 kilometres of fibres between the 212 construction phase, other (sub-)millimetre
Gie Han Tan 2 antenna pads and the correlator located telescopes have also benefitted from
Ciska Kemper 2 in the Array Operations Site (AOS) techni- these advances and have since overtaken
Neil Phillips 2 cal building at an elevation of 5100 metres. ALMA, especially in terms of the instanta-
During the construction phase, the Uni- neous bandwidth that can be observed.
versity of Bordeaux (UB) was responsible To keep ALMA at the forefront of science
1
LAB, University of Bordeaux, France for delivering two subsystems: the digitiser and technology, the ALMA2030 Develop-
2
ESO and the Tunable Filter Bank (TFB). ment Roadmap (Carpenter et al., 2019)
both defined new key science drivers and
The digitiser module was built on two prioritised increasing the overall instanta-
The Atacama Large Millimeter/submi application-specific integrated circuits neous bandwidth of ALMA by a factor of
llimeter Array (ALMA) comprises 66 an- (ASICs), a 3-bit analogue-to-digital at least two (but preferably four). This
tennas working as a powerful interfer- converter (ADC) and a demultiplexer, entails increasing the IF bandwidth of the
ometer. High-speed digitisation, signal designed by UB with a bipolar comple- receivers and the associated electronics
transmission over several tens of kilo- mentary metal-oxide-semiconductor and correlator. With these improvements,
metres from the receivers to the corre- (BiCMOS) technology from ST-Microelec- ALMA will be 2–4 times faster for spectral
lator, and complex data processing all tronics (see Baudry et al., 2006). This line surveys such as those required to
require state-of-the-art technologies. module performs the signal digitisation in study the chemical complexity in star-
The ALMA2030 Development Roadmap the 2–4 GHz intermediate frequency (IF) forming regions, but also when it is used
calls for an increase in the bandwidth bandsa of two polarisations resulting from as a redshift machine (see, for example,
by at least a factor of two, implying a the first and second analogue frequency Reuter et al., 2020). All continuum pro-
major upgrade of the entire signal path. conversions, with the digitisers sampling jects will at the same time benefit from
We present here the results of a detailed in the second Nyquist zone. The receiver the improved sensitivity and imaging
study looking at how to upgrade the usually produces an IF of 8 GHz, so each fidelity that comes with such a dramatic
ALMA digital system, including digitisa- antenna is equipped with four digitisation increase in bandwidth.
tion, data pre-processing, and data modules.
transmission to cope with bandwidths Such a major upgrade is of course more
more than four times the current ones. The TFB, part of the correlator, is a digi- complex for an interferometer with as
At the same time, this system will con- tal electronic system based on field- many as 66 antennas. Whilst most of the
tribute to increasing the nominal corre- programmable gate array (FPGA) tech- development efforts have thus far concen-
lation efficiency from 88% to 99%, and nology, clocked at 125 MHz, that per- trated on completing the original receiver
prepare ALMA for longer baselines of forms massively parallel data processing. complement, this effort will come to an
up to 100 kilometres. Each TFB card can process one of the end in a few years now that the final
2-GHz basebands generated by the receiver (Band 2) is on track to be built
digitiser module to extract up to 32 sub- by an international consortium led by ESO
Introduction bands of 62.5 or 31.25 MHz, on which (Yagoubov et al., 2020). This allows ALMA
the correlation is performed. The TFB to now take on the ALMA2030 bandwidth
ALMA is by far the most powerful (sub-) implements a spectral selection and upgrade, involving receivers, digitisers
millimetre interferometer ever built, and zoom effect, and increases the total num- and the correlator. During the last dec-
has produced transformational science ber of spectral channels produced by ade, the Laboratoire d’Astrophysique de
results throughout its first decade of opera- the correlator by a factor of 30 (7680 per Bordeaux (LAB) has completed a set of
Digital system upgrade

proposed by LAB Correlator
Back-end Emitter Receptor
DWDM
Transport fibre
Receiver Digitiser DTX Optical Optical

DRX/TFB
amplifier amplifier
~ 100 km
Multiplier Demultiplier
Figure 1. Upgrade of
(0–20) GHz IF 40 GS s–1
960 Gb s –1
the digital system as
* 2 polar * 6 bits proposed by the LAB,
* 2 side bands * 2 polar University of Bordeaux
* 2 side bands (both AOS and OSF
Array Operations Site (AOS) Operations Support Facility (OSF) locations are considered
for the new correlator).

development studies supported by ESO Figure 2. Test bench
of the Micram digitiser
to identify appropriate solutions for the
module in the technical
digitiser, digital signal processing and facility at the LAB,
data transport that enable meeting these University of Bordeaux.
ambitious new goals. In addition to the The digitiser perfor-
mance is measured at
bandwidth extension, the more modern
40 GS s –1 using the Intel
ADCs also allow one to increase the nomi- Stratix 10 GX FPGA
nal correlation efficiency from the current development kit.
88% to almost 99%, which in terms of
sensitivity is equivalent to adding 6–8 new
antennas to the array, even for single-line
observations. The advanced digitiser
solutions studied by the LAB match well
the aims of the future ALMA system archi-
tecture using only a single, very wide-band
IF stage instead of the dual heterodyne
conversion approach currently in use. ALMA digital system, from digitisation to cal fibre. Removing most of the analogue
correlation (see Figure 1). parts, currently used for the second fre-
This single frequency conversion archi- quency conversion prior to digitisation,
tecture has several major advantages for we would enhance back-end versatility,
the user: Digitisation reproducibility and reliability, and ease
– There is no loss of instantaneous band- calibration and failure analysis.
width arising from the use of multiple, Digitisation is one of the main techno
staggered, non-ideal IF pass bands. In logical challenges in respect of increasing Today the baseline solution for the
the current system 4 adjacent bands of the ALMA instantaneous bandwidth ALMA2030 digitisation is an ADC module
2 GHz each are used; however, the because the market for ADCs with ana- from Micram Microelectronic GmbH.
effective total bandwidth is only 7.5 GHz. logue bandwidths and sampling frequen- This commercial device was initially speci-
– The lack of discontinuities in the band- cies above 10 GHz is extremely small fied for 34 GS s –1, but has been exten-
pass curves of adjacent filters, since and unsteady. For the past ten years the sively evaluated by the LAB group up to
the whole IF range will be covered by electronics group at the LAB has been 40 GS s –1, and subsequently also by
one passband, makes calibration for investigating homemade fully custom ASIC Micram. The Micram digitiser chip is built
this anomaly redundant. solutions, based on CMOS or BiCMOS on two internally interleaved ADC cores.
– There is no longer a need for oscillators technologies, continuing the development Complex calibration is required to opti-
to convert the astronomical signal from work it undertook for the construction mise the linearity of each core and mini-
the IF1 range to the IF2 range. In prac- phase, and has in parallel carried out an mise the mismatch between cores. With
tice these strong oscillator signals pro- ongoing survey of the solutions announced this solution we can achieve direct digiti-
duce unwanted spurious, ghost signals or under development by small micro- sation of, for example, 2–19.5 GHz IF at
in the astronomical observations. electronics startups and global electron- 40 GS s –1, ten times better than the cur-
ics companies. Several of the most rent digitiser, and we also significantly
The main objective of the first LAB study promising solutions have been evaluated increase the quantisation efficiency since
in 2013 was to evaluate technologies in the laboratory to measure their actual the Micram device is a 6-bit ADC (as
which could be used within a ten-year performance and assess the complexity against 3 bits for the current digitiser).
timeframe to upgrade the sub-systems of their implementation and potential use
originally delivered by the LAB. The sec- for ALMA. Various digitisation topologies Digitiser requirements
ond study performed by the LAB was (for example, multi-rate, interleaved) have
focused on the identification and the also been considered to achieve digitisa- The requirements that drive the selection
evaluation of the critical devices required tion of a bandwidth larger than half the of the digitiser are the sampling frequency,
to upgrade digitisation, data transmis- sampling frequency, the limit given by the the bandwidth and the quantisation effi-
sion, and digital signal processing, in Shannon sampling theorem (see, for ciency. Indeed, the maximal sampling fre-
accordance with the general ALMA2030 example, Tan & Jiang, 2019). quency (combined with the digitisation
roadmap, which had been progressively topology adopted) and the digitiser band-
defined in the meanwhile. This second Our primary objective has always been to width together define the effective IF
study also included system architecture design a digital back-end including the band which can be digitised, and thus
considerations and comparisons. The digitisation, data transmission, and possi- the instrument’s instantaneous bandwidth.
most recent LAB study, which is ending bly some digital signal processing func- Moreover, the quantisation efficiency is
in June 2021, aims to confirm these criti- tions, which would be able to digitise the a direct contributor to the overall instru-
cal choices regarding the ALMA2030 full IF bandwidth delivered by the new ment sensitivity. If the question of the
specifications which are now available. generation of receivers and format the bandwidth and sampling frequency is
All three studies focus on upgrading the digital samples for transmission over opti- relatively straightforward, the question

Instrumentation Quertier, B. et al., Upgrade Strategies for the ALMA Digital System
0 100 Figure 3. The Micram

digitiser’s bandwidth
ADC2 freqresp 99 ADC2 effmin
(measured using a
–1 sine wave input; note
98 that only frequencies
97 below 20 GHz would
Amplitude (dBFs)
–2 be used when sampling
Efficiency (%)
96 the first Nyquist zone at
40 GS s –1) and quantisa-
–3 95 tion efficiency (measured
using a Gaussian noise
94
input that includes a
–4 93 passband gain variation
of 5.4 dB over 16 GHz).
92
–5
91
–6 90
0 10 20 30 40 50 –10 –5 0 5 10
Input signal frequency (GHz) Noise input power (dB)
of the quantisation efficiency, and how ADC from SNR measurements instead of ple of 100 devices. It was also demon-
to define and measure it, is much more over-constraining the digitiser specification strated that the bandwidth at 40 GS s –1 is
complex. using the effective number of bits (ENOB) beyond 20 GHz (see Figure 3), that the
as is often suggested. Indeed, it would digital sample capture and synchronisation
Quantisation efficiency is the relative loss be particularly unfortunate to be mistaken with an FPGA was suitable for long-term
in signal-to-noise ratio (SNR) resulting in the digitiser evaluation and selection applications, that the mismatch calibration
from the quantisation process; it depends process, considering that there are very process of the two cores appears effi-
on the number of quantisation levels and few candidates that are capable of cient at rejecting the ghost image arising
the statistical property of the input signal. achieving the ALMA2030 objectives. from interleaving below the quantisation
Thompson, Emerson & Schwab (2007) noise floor, and that the threshold response
provide a method for calculating the Performance of the Micram solution estimated from Gaussian noise histograms
quantisation efficiency for any number of makes it possible to optimise the ADC
uniformly spaced levels as a function of We ran four test campaigns with the linearity for radio astronomy. Finally, we
the level spacing, using mathematical for- Micram device over the past 3 years. The have been able to estimate the quantisa-
mulas with radio-astronomy-like signal first two were held at the Micram facility tion efficiency from SNR measurements
(Gaussian for a perfect ADC). The devia- and allowed us to measure the general at around 98% for standard astronomical
tion from the ideal quantisation step width, performance, the stability over time and observations (passband gain variations
which necessarily exists with a real ADC the temperature of the digitiser module < 5.4 dB over 16 GHz and signal level
because of manufacturing imperfections, operating at 34 GS s –1, on noise signals. changes ≤ 4 dB).
is called differential nonlinearity. Appro- Two additional test campaigns were then
priate calibration is needed to minimise undertaken in the LAB technical facility
this nonlinearity, but some degradation of (see Figure 2), based on our own digital
the quantisation efficiency is unavoidable. signal processing and calibration proce-
In our latest study we describe a novel dures, when we were able to demonstrate Figure 4. Possible data transmission scheme from
approach which consists of estimating the the performance up to 40 GS s –1. This antennas to correlator (both AOS and OSF locations
effective quantisation efficiency of a real was later confirmed by Micram on a sam- are considered).
Antenna AOS OSF

Emitter Receptor
Transport fibre
DG DSP/OT Optical Optical

DSP/OT Correlator DSP/OT Correlator
amplifier amplifier
~ 100 km
66 × DWDM
12ch 10 G
DCP-M40-PAM4-ER
–
8 × DWDM
33ch 400 G – Automatic fibre distance measurement and
Example of a 400 Gb s –1 Data Centre
Emitter Receptor
dispersion compensation setting
Interconnects (DCI) optical TRX 66 × 4 × 400 G Transport fibre – Complete zero-touch automatic optical setup,
Inphi Inc., Model IN-Q3JZ1-TC just like using a passive multiplexer
Optical Optical
Modulation: PAM-4 – Optical performance

amplifier amplifier
~ 100 km
Transmission reach up to 120 km @ 400 G Multiplier Demultiplier

– DCP-M40-PAM4-ER: 0–40 km, 0–14 dB

Digital data pre-processing and data gated by the LAB group over the past because at each stage we need many
transmission 5 years, in order to propose an upgrade to high-speed FPGA transceivers to capture,
our initial contribution with the TFB. Note transmit, and receive the extended data
The antennas are connected to the that the current baseline correlator archi- flow. Since this kind of FPGA will neces-
AOS technical building by a network of tecture is a digital hybrid XF design (or sarily come with a very large number of
optical fibre cables. Each antenna is FXF). However, when frequency division of computational resources, we can con-
connected by eight single-mode optical the input baseband is bypassed, then the sider various partitioning configurations
fibres, allocated as follows: one fibre for correlator behaves as a pure XF system; of the overall signal processing between
the Data Transmission System, one fibre both operating modes (frequency-division the antenna, the AOS and the OSF build-
for the Photonic Local Oscillator and low- mode and time-division mode) are offered ings in order to minimise the cost, ease
frequency timing references, two fibres to users. This brings greater flexibility and the deployment, and increase the versa-
for the monitoring and control communi- makes it possible to have an independent tility of the system. For example, the
cation, and four spare fibres. The data zooming factor within different spectral pre-processing and data transmission
flow generated by each antenna is trans- windows (a tradeoff between effective functions of the current system could be
mitted over a single optical fibre using bandwidth and spectral resolution). merged in a single FPGA. Moreover,
wavelength division multiplexing, a tech- having a first DSP stage like a frequency
nology which multiplexes a number of Data transport from antenna to OSF division at the antenna would be the most
optical carrier signals onto a single optical via AOS straightforward strategy to transmit the
fibre by using different wavelengths (i.e., overall data stream over the optical system
colours) of laser light. The current system One of the ALMA2030 priorities gets because of considerations concerning
is based on 20-channel dense wavelength implemented by having extended base- the detailed electrical interface of the
division multiplexing (DWDM) with10 G lines. The cable length to such remote transceivers and impacts on frame syn-
per channel, where only 12 channels are new antenna pads may be significantly chronisation. Figure 5 gives an overview
populated. The current throughput of longer than the baseline length to the of one of the possible architectures we
120 Gb s –1 per antenna could then be AOS, so lengths of 50 or even 60 kilo- have identified to perform the frequency
increased to 200 Gb s –1, but this would metres cannot be excluded at this point. division prior to the correlation. Here, over-
not in any case be sufficient to support The 400G ZR data centre interconnects sampled polyphase filter banks are used
the IF bandwidth doubling that is a mini- (DCI) technology allows for transmission to provide the ability to switch between
mum objective for ALMA 2030. reach up to 120 kilometres. The Inphi time-division mode and frequency-division
Corporation, a leader in high-speed data mode while keeping the same usable
The location of the new-generation corre- movement interconnects, announced in bandwidth and throughput. One important
lator will almost certainly be at the Opera- March 2021 the commercial availability conclusion is that the transceiver require-
tion Support Facility (OSF) and no longer and production ramp-up of COLORZ II ments and DSP partitioning are compati-
at the AOS building. This preference has 400ZR, the industry’s first quad small- ble with the idea of a generic FPGA
been expressed by several key stake- form pluggable double-density coherent board (DSP/OT, as seen in Figure 4) that
holders over the past years, for example transceivers for cloud DCIs. It would be could be used at both ends of the optical
at the ALMA 2030 workshops organised the perfect solution for the antenna-to- data transmission system, to implement
by the US National Radio Astronomy AOS links. The distance between the various functions required along the data
Observatory (NRAO) and the National AOS and the OSF is less than 40 kilo- path. The only differences would be the
Astronomical Observatory of Japan metres. 48 single-mode optical fibres firmware to be implemented in the generic
(NAOJ) in 2020. For the upgrade design are available for this connection. Installing FPGA board. This would be a tremendous
there is a near consensus within the additional fibres might not be possible advantage in terms of validation, manu-
ALMA developer community that the FFX so we have identified a fully commercial facturing, fault analysis and maintenance.
architecture (where the first F indicates a solution for a complete 40-channel mul Using these FPGAs to perform the 1st F, at
frequency division of the signal, the sec- tiplex/demultiplex DWDM module com- either end of the DTS link, rather than hav-
ond F stands for a Fourier transform stage patible with 1/10/40/100/200/400G ing additional FPGAs in the correlator for
and the X represents the cross-correlation Ethernet. Eight of these modules would this task, could considerably improve the
stage of the signal processing) would be increase the ALMA optical transport cost and power efficiency of ALMA2030.
preferred, considering the possibilities capacity to 2 Tb s –1 per antenna, while
offered by the wide multipliers in digital using only eight single-mode fibres. This
signal processing (DSP) hardware such as solution, combined with 400G ZR trans- Outlook: prototyping and on-site
FPGA and the finest spectral resolution ceivers, allows for antenna-to-AOS dis- demonstration
target of 1 kHz. Baudry et al. (2017) give tances of up to 80 kilometres.
several examples of high-spectral-resolu- Our current plan for this upgrade is
tion architecture where a frequency divi- Generic board DSP/optical transmission described by the global block diagram in
sion of the input baseband is followed by for the “1st F” and data transmission Figure 6, with on-site demonstration in
a pure FX correlator. This pre-processing, potentially three steps to ensure compati-
often just called “the 1st F” in the DSP For ALMA2030, state-of-the-art FPGAs are bility with the existing infrastructure (physi-
architecture, has been extensively investi- required at both ends of the optical fibre, cal locations, power supplies, reference

Instrumentation Quertier, B. et al., Upgrade Strategies for the ALMA Digital System
20 GHz 8 × 2.5 GHz 8 × 2.5 GHz 64 × 312.5 MHz 128 × 312.5 MHz 128 × SB
Real part Digital Real part Decimation

Polyphase
1 Pol 6 bits 3–6 bits filter 6 bits mixer filter Complex/ 2–6 bits
Polyphase Optical Real
× DFT DFT SEL x
filter transmission converter 62.5–312.5 MS s –1
USB/LSB 40 GS s
–1
5 GS s –1 625 MS s –1
Real input Polyphase Decimation Real output
Complex Complex DDS
Imag. part filter Imag. part filter
Control
16 GHz 2 GHz 2 GHz
0 5 10 15 20 GHz 0 625 1250 1875 2500 GHz 0 5 156.25 312.5 GHz

or or
or
2 GHz
2 GHz
8 GHz
0 5 10 15 20 GHz 0 625 1250 1875 2500 GHz 0 156.25 312.5 GHz
Figure 5. (Above) Example of DSP architecture.
Figure 6. (Left) In black:

the current system
showing the digitiszer
module and the Tunable
Emitter Receptor Filter Bank, two sub-
≤ 200 Gb s –1
systems of the ALMA
Transport fibre
instrument designed
IFDC DG DTX DRX TFB Correlator and delivered by the
LAB, University of
Optical Optical
amplifier amplifier
Multiplier
~ 100 km
Demultiplier
Bordeaux.; In colour: the
digital system upgrade
plans proposed by the
1 2
Emitter Receptor
LAB, Université de
≥ 400 Gb s –1 B ordeaux, with on-site
Transport fibre
demonstration in three
IFDC potential steps before
DG DRX Correlator
DTX a full production and
3
Optical Optical
amplifier amplifier
~ 100 km
integration.
Step Quantisation Efficiency Bandwidth Spectral coverage
#1 3 bits × 2 bits 0.84 2 × 4 GHz/pol 128 × 62.5 MHz/pol
#12 6 bits × 2 bits 0.88 2× 16 GHz/pol 128 × 62.5 MHz/pol
#13 6 bits × 6 bits 0.99 2× 16 GHz/pol 256 × 125 MHz/pol
signals, control and monitoring interfaces Acknowledgements Tan, L. & Jiang, J. 2019, Digital Signal Processing,
(Cambridge, Massachusetts: Academic Press),
and optical fibres) before a full production
We would like to acknowledge the Micram and Intel 13 et seq.
and integration. The recommended strat- companies for their technical support and various Thompson, A. R., Emerson, D. T. & Schwab, F. R.
egy is to deploy the new ALMA 2030 dig- hardware loan agreements which made this work 2007, Radio Science, 42
ital system in parallel with existing hard- possible.
ware, with all new components installed
Notes
at different physical locations, so that the References
current system can continue science a
The term “IF band” is used to distinguish the
operations until the new system is com- Kemper, C. 2020, The Messenger, 180, 42 on-sky, radio frequency (RF) band (in the
Baudry, A. et al. 2006, The Messenger, 125, 37 35–950 GHz range for ALMA) from that accessible
missioned. The bulk of commissioning
Baudry, A. et al. 2017, ALMA Memo, 607 to the digitisation electronics (typically < 20 GHz).
could then be accomplished during peri- Carpenter, J. et al. 2019, The ALMA Development
ods of unfavourable or less favourable Roadmap, arXiv:1902.02856
observing conditions (i.e., in daytime), Reuter, C. et al. 2020, ApJ, 902, 78
Yagoubov, P. et al. 2020, A&A, 634, A46
while science operations continue with
the current system during optimal
conditions.

Astronomical Science
This quartet of galaxies makes up the Hickson

Compact Group HCG 86. It was observed with
ESO’s VLT Survey Telescope (VST) as part of the
VST Early-type Galaxy Survey (VEGAS) large
p rogramme. Because of their compactness, such
ESO/Ragusa, Spavone et al.
groups are ideal environments to study galactic

interactions, which can sometimes lead to galax-
ies merging with each other. Not all bright objects
in this image belong to HCG 86. The four members
of HCG 86 are seen from Earth as arranged in
triangular shape, with three of them on a straight
line and one to the left; the bright objects below
of the elongated galaxy are not part of the quartet.
Astronomical Science DOI: 10.18727/0722-6691/5241
The INvestigate Stellar Population In RElics (INSPIRE)

Project — Scientific Goals and Survey Design
Chiara Spiniello 1, 2 compact galaxy with a size a factor of

Crescenzo Tortora 2 ~ 4 smaller than local massive galaxies
Giuseppe D’Ago 3 (the so-called “red nuggets”). Then a
Nicola R. Napolitano 5, 2 second, prolonged phase, dominated by
and the INSPIRE Team mergers and gas inflows, is responsible
for the structural evolution and size growth
from z ~ 1 to today. The ultra-compact
1
Sub-Department of Astrophysics, retaining the memory of the star formation objects formed at high redshift likely end
Department of Physics, University of activity that occurred in the earlier phases up forming the cores of giant local galax-
Oxford, UK of galaxy formation. Secondly, they ies, whilst subsequently accreted or
2
INAF – Astronomical Observatory of account for more than half of the total newly formed stars remain preferentially
Capodimonte, Naples, Italy stellar mass in the current Universe and in the external regions. Unfortunately, this
3
Institute of Astrophysics, Pontificia are responsible for most of its chemical “accreted” material overlaps, along the
Universidad Católica de Chile, Santiago, enrichment. Understanding the history line of sight, with the spatial and orbital
Chile of the assembly of the most massive gal- distributions of the in-situ pristine light
5
School for Physics and Astronomy, Sun axies throughout cosmic time is therefore that encodes the information about
Yat-sen University, Guangzhou, China crucial to constraining models of galaxy high-z baryonic processes, irreversibly
formation and evolution. limiting our resolving power and hamper-
ing our ability to study the early phases
Relics are the ancient fossils of the early The first generation of massive ETGs is of galaxy formation. Luckily, since merg-
Universe. They are ultra-compact and already in place at z ~ 3, i.e., only two ing is a stochastic phenomenon, a small
massive galaxies that formed only a few billion years after the Big Bang. However, fraction of red nuggets survives intact
(1–2) billion years after the Big Bang, in at this redshift, most of the massive red until the present day, without experienc-
a short and intense burst of star forma- objects are found to be 3–5 times smaller ing any merger or interaction and thus
tion, and then evolved passively and in size than their counterparts in the local remaining as massive and ultra-compact
undisturbed until the present day, com- Universe, and thus they are 30–100 times as they formed: relic galaxies (Trujillo et
pletely missing the accretion phase denser (van Dokkum et al., 2008). To al., 2009). Since a detailed study of the
predicted for the assembly of local reconcile these observations, a two-phase stellar populations of high-z red nuggets
giant early-type galaxies. As such, they formation scenario (Oser et al., 2010) has would require prohibitive integration times
represent a unique opportunity to put been proposed to explain the mass with the currently available facilities, relics
precise constraints on the first phase of assembly and evolution across cosmic are the only systems that allow us to
structure formation in the Universe. time of very massive galaxies (Figure 1). study the physical processes that shaped
Since the number of relics predicted at the mass assembly of massive galaxies in
each redshift depends heavily on the Initially, a series of intense, fast, dissipative the high-z Universe in the amount of
mechanisms responsible for the accre- processes forms their central “bulk” mass detail currently reachable only for the
tion and growth of massive galaxies, (at z > 2) generating, after star formation nearby Universe.
obtaining number counts at 0 < z < 0.5 quenches, a massive, passive and very
is a very powerful way to v alidate and
disentangle different possible physical
First phase
scenarios driving their formation and
size-evolution. INvestigating Stellar Relic
galaxy
Population In RElics (INSPIRE) is an
ongoing project based on an approved Starburst Red
ESO Large Programme, targeting Big Bang galaxy nugget
52 relic candidates with the X-shooter
spectrograph at ESO’s Very Large Tele-
scope with the aim of building the first
statistically large catalogue of relics at 5 billion
Today
0.1 < z < 0.5. years
1 billion 3 billion
Relic galaxies: fossils of the ancient years years
Universe
Massive early-type galaxies (ETGs) play a Mergers and Local giant

crucial role in the context of cosmic interactions early-type
structure formation and mass assembly. Figure 1. A sketch of the two-phase formation sce-
galaxy
Firstly, they contain the oldest stars at nario for the mass assembly and cosmic evolution of
any epoch of the Universe’s history, thus massive early-type galaxies. Second phase (accretion)

2.5 2.0
Szomoru, Marijn & van Dokkum (2012)
Normal-sized galaxies (KiDS)
1.5
2.0
log10 (Re [kpc])

1.0
(g–i)
1.5 0.5
0.0
INSPIRE DR1
1.0 INSPIRE next releases
0.1 0.2 0.3 0.4 10.8 11.0 11.2 11.4 11.6

Redshift log10 (M /M๬
*
But how many relics exist in the Universe spectroscopic follow-ups at many differ- Figure 2. Left: Observer-frame (g–i) colour v ersus
redshift for the 52 INSPIRE objects (red circles). Filled
at each redshift, and how many survive ent ESO and non-ESO telescopes have
symbols indicate the 19 galaxies included in DR1,
until the present day? How can they pas- been obtained to validate these objects, while empty ones are those that will be released in
sively evolve through cosmic time without inferring their redshifts and masses and the future. The grey region shows the predicted
experiencing any interaction? What can thus confirming them as ultra-compact values for single stellar population models with ages
of 8 Gyr (black) or older. Right: Stellar mass–size
we learn about the merging history and massive galaxies (Tortora et al., 2018;
relation of relic candidates (same colour code as
the evolution of the most massive ETGs? Scognamiglio et al., 2020). These objects above) compared to the relation for normal-sized
Is there a physical scenario, predicted are the perfect relic candidates, as they passive galaxies from the KiDS survey, shown as
by hydrodynamical cosmological simula- all have large stellar masses (M∗ > 6 × a dark-grey shaded region and to high-redshift red
nuggets from the collection of Szomoru, Marijn &
tions, that is able to explain all the current 1010 M☉), small sizes (Re < 2 kpc) and
van Dokkum (2012).
observational results at all redshifts? very red colours indicative of an evolved,
These and other questions remain to be old stellar population (see Figure 2).
answered and constitute the main topics after S21a), we presented a pilot pro-
of the INvestigate Stellar Populations In The INSPIRE observing strategy has gramme demonstrating the feasibility of
RElics (INSPIRE) Project. been optimised to capitalise on relatively the project and showing the kinematical
sub-standard observing conditions (see- and stellar population analysis for three
ing up to 1.2 arcseconds, clear nights, objects representative of the whole sam-
The INSPIRE ESO Large Programme: grey lunar phase with a lunar illumination ple. Results for the first 19 galaxies, com-
observational strategy and current status fraction up to 0.5), allowing the objects pletely observed before the end of 2020,
to be easily scheduled into the observing are presented in Spiniello et al. (2021b,
The INSPIRE1 ESO Large Programme queue. Moreover, the selected targets hereafter S21b), and released as part of
(ID: 1104.B-0370, PI: Chiara Spiniello) span a very wide range in right ascension the first INSPIRE data release (DR1). DR1
was approved two years ago to spectro- and declination, with an optimum observ- was made publicly available through the
scopically follow up with the X-shooter ing time spread over the full year. We also ESO Archive Science Portal4 in March
spectrograph 52 ultra-compact, red, mas- note that we have many systems with 2021. It comprises one-dimensional (1D)
sive galaxies with redshifts 0.1 < z < 0.5. declination below –30 degrees, perfect spectra in the UVB and VIS arms of
They are the final products of a dedicated as “fillers” on nights with strong wind com- X-shooter for 19 ultra-compact massive
project within the Kilo Degree Survey2 ing from the north. The integration time galaxies that were observed between
(KiDS) collaboration, one of the three on target, which varies from target to 22 October 2019 and 31 December 2020
ESO VST Imaging Public Surveys3. This target according to their r-band luminosity and were classified as relics or not in
project started in 2015 with the goal of and surface brightness, has been set to S21b. Near-infrared spectra, which are
building the largest catalogue of photo- reach a signal-to-noise ratio sufficient to not necessary for the confirmation of the
metrically selected passive (non-star- infer the age, metallicity and α-abundance relic nature of the systems but which are
forming) galaxies with incredibly small of the stellar populations. crucial to correctly inferring the stellar ini-
sizes but stellar masses comparable to tial mass function (IMF), will be analysed
those of massive local ETGs (Tortora At the time of writing, data on 38 out of and released in a separate forthcoming
et al., 2016), and with secure structural 52 galaxies have been collected, and data release, accompanied by a dedicated
parameters inferred from gri-band KiDS 36 objects have already been completely scientific publication (see next section for
images (Roy et al., 2018). Subsequently, observed. In Spiniello et al. (2021a, here- more details).

Astronomical Science Spiniello , C. et al., The INvestigate Stellar Population In RElics (INSPIRE) Project
Figure 3. Four objects from INSPIRE

J0847+0112 J0316-2953 J0857-0108 J2327-3312 DR1, each with a different “degree
of relicness’’. In the green box, we
show the most extreme and the least
extreme relics, whilst in the red box we
show the non-relics with the least and
most extended star formation history.
For each galaxy, the top panel shows
a 10 × 10 arcsecond colour-combined
(gri) KiDS image, the middle panel
1.00 1.00 1.00 1.00
Cumulative mass %
Cumulative mass %
Cumulative mass %
Cumulative mass %
shows the fraction of stellar mass
0.75 0.75 0.75 0.75
assembled in cosmic time since the
0.50 0.50 0.50 0.50
Big Bang, varying the [Mg/Fe] by ±0.1
0.25 0.25 0.25 0.25
(grey area), and the bottom panel
0.00 0.00 0.00 0.00 shows a zoom-in to region of the
0 2 4 6 8 10 12 0 2 4 6 8 10 12 0 2 4 6 8 10 12 0 2 4 6 8 10 12
Time since Bing Bang (Gyr) Time since Bing Bang (Gyr) Time since Bing Bang (Gyr) Time since Bing Bang (Gyr) spectrum used to compute the veloc-
ity dispersion, and the stellar age and
Flux (10 –17 erg s –1 cm –2 Å –1)
Flux (10 –17 erg s –1 cm –2 Å –1)
Flux (10 –17 erg s –1 cm –2 Å –1)
Flux (10 –17 erg s –1 cm –2 Å –1)

6.0 4.0 metallicity. The values for the stellar
4.0 4.0 velocity dispersion, the mass assem-
4.0
2.0 bled at the end of the first formation
2.0 2.0
2.0
phase (z ~ 2, 3 billion years after the
0.0 0.0 0.0
Big Bang) and the time of final assem-
4000 5000 6000 7000 4000 5000 6000 7000 4000 5000 6000 7000 4000 5000 6000 7000 bly (when more than 95% of the stellar
Restframed wavelenght (Å) Restframed wavelenght (Å) Restframed wavelenght (Å) Restframed wavelenght (Å) mass was in place) are also provided.
σ ∼ 220 km s –1 σ ∼ 170 km s –1 σ ∼ 145 km s –1 σ ∼ 215 km s –1
* * * *
M , 3 Gyr = 100% M , 3 Gyr = 87% M , 3 Gyr = 79.5% M , 3 Gyr = 27.5%
* * * *
tfin = 1.75 Gyr tfin = 6.5 Gyr tfin = 4.5 Gyr tfin = 9 Gyr
INSPIRE DR1: kinematics, stellar popu- overall larger σ in relics would indicate a The INSPIRE DR1 catalogue of 10 relics
*
lation analysis and relic confirmation for higher dynamical mass compared to non- known to date augments by a factor of
the first 19 systems relic galaxies of similar stellar mass. This 3.3 the total number of confirmed relics,
can have two possible physical explana- also enlarging the redshift window out-
Here we focus on three main scientific tions. On the one hand, relics could have side the local Universe (up to z ~ 0.4). It
results, obtained from the 19 systems a larger number of stars with very low is therefore the largest publicly available
with complete observations until the end masses — i.e., a dwarf rich IMF — which collection at the moment. Thanks to the
of 2020, presented in S21b (also includ- contribute very little to the light but sub- larger number of systems, we can con-
ing the three pilot galaxies presented stantially to the dynamical mass. On the firm the existence of a “degree of relic-
in S21a): i) the integrated velocity disper- other hand, relics could have a larger ness” (Figure 3), already hinted at in the
sion values inferred from the 1D spectra fraction of dark matter in the central literature, quantifying how fast the star
encapsulating 50% of the light (30% in regions or, possibly, dark matter halos formation histories are. This degree of
S21a) of the galaxies and those inferred with a higher central mass density. relicness might correlate with the struc-
from optimally extracted 1D spectra; ii) a tural parameters of the galaxies (for exam-
precise estimate of the stellar age, metal- Spectroscopic ages are very old for ple, size or ellipticity), dynamical proper-
licity and [α/Fe] abundance of the stellar 13 out of 19 galaxies, in agreement with ties (i.e., the extreme relics are the
populations, from line indices and full the photometric ones, and metallicities are systems with the largest velocity disper-
spectral fitting; and iii) star formation his- almost always (18 out of 19) super-solar, sions) and/or with the environments in
tories and “relic confirmation” based confirming the mass-metallicity relation. which they live. We will investigate this
on the fraction of stellar mass assembled The [Mg/Fe] ratio, that sets the clock of matter further once the full INSPIRE sam-
during the first phase of the two-phase the star formation (i.e., larger values indi- ple has been analysed, as described in
formation scenario and on the cosmic cate a faster star formation episode) is the next section.
time of “final assembly”. also larger than solar for the great major-
ity of the galaxies, as expected.
The velocity dispersion values are always Future INSPIRE data releases and final
relatively high (σ > 150 km s –1), confirm- We find that 10 objects have formed scientific goals
*
ing the massive nature of the systems. more than 75% of their stellar mass within
Although with limited statistics, we observe 3 Gyr of the Big Bang, when the first At least two further data releases will be
that the galaxies confirmed as relics have phase is believed to end, and therefore issued for INSPIRE. In the second data
a systematically larger velocity dispersion we classify them as relics. Amongst release, foreseen in roughly six months,
compared to non-relics, for the same these, we identify 4 galaxies which had we will add the near-infrared (NIR) spec-
stellar mass. Since σ is approximately already fully assembled by that time. tra of the 19 systems released in DR1.
*
proportional to the dynamical mass of the They are therefore “extreme relics” of the Extending the wavelength range towards
galaxy within the half-light radius, the ancient Universe. the red, we will be able to constrain, at

least for the stacked spectrum of all the logue of relics at different redshifts is a able to compute the ratio between stellar
10 relics, the slope of the IMF at the low- conditio-sine-qua-non to disentangle pos- and dynamical mass (M /Mdyn) which is
*
mass end. The stellar IMF is the distribu- sible physical scenarios driving the for- a valuable tracer of the likely mechanism
tion of stellar masses that form in one mation and size-evolution of galaxies. At by which galaxies grow. In fact, under
star-formation event in a given volume of the same time, the characterisation of merger-driven growth, the M /Mdyn ratio,
*
space. Hence, all the observable proper- the structural, dynamical, environmental, measured within the effective radius,
ties of stellar systems are heavily influ- and stellar properties of the non-relics, should decrease with time; this decrease
enced by their IMFs, since the mass of a ultra-compact systems that have just would be stronger in the case of minor
star determines its subsequent evolution- stopped forming stars, will provide a key mergers. Thus, under the hypothesis that
ary path. Furthermore, almost every test for models of galaxy formation. In relics experienced very few mergers, or
observable property of a galaxy depends fact, according to hydrodynamical simula- none at all, they should all show a very
on its mass-to-light ratio (M/L) and the tions very compact and very massive gal- high M /Mdyn, compared to normal-sized
*
low-mass end of the stellar IMF is crucial axies should only form in gas-rich envi- galaxies of similar stellar masses, and
in determining that. In fact low-mass stars ronments, and thus early on in time, and thus a smaller amount of dark matter.
account for more than half of the mass then undergo a series of mergers and
budget but they contribute very little to interactions which should cause a growth
the integrated luminosity of a galaxy with in size. We will compare the INSPIRE gal- Acknowledgements
an old stellar population. This makes axies (both relics and non-relics) with The authors wish to thank the ESO Archive Science
the characterisation of the low-mass IMF ultra-compact simulated galaxies from Group, led by Magda Arnaboldi, for the precious
slope from integrated light a very difficult different hydrodynamical simulations to help in making the INSPIRE DR1 publicly available as
task but also a crucial ingredient for par shed light on the formation scenario Phase 3 Collection. CS is supported by a Hintze Fel-
lowship at the Hintze Centre for Astrophysical Sur-
titioning galaxy mass into stellar and dark and mass assembly of these extremely veys at the University of Oxford, which is funded
matter, and for understanding how they compact, passive and massive objects, through generous support from the Hintze Family
interact in the internal region of galaxies. and to understand the conditions that Charitable Foundation. NRN acknowledges financial
This, in turn, allows one to predict the allowed these objects to survive intact to support from the “One hundred top talent program
of Sun Yat-sen University” grant N. 71000-18841229.
luminosity evolution of passively evolving the present day without undergoing any GD acknowledges support from CONICYT project
stellar systems and thus to correctly accretion event. Basal AFB-170002.
interpret the cosmic evolution of the most
massive galaxies in the Universe. Likewise, we will compare the final
References
INSPIRE sample with normal-sized ETGs
The third data release, which will be at similar redshifts and with similar struc- Oser, L. et al. 2010, ApJ, 725, 2312
made available upon completion of all the tural and photometric (magnitudes and Roy, N. et al. 2018, MNRAS, 480, 1057
spectroscopic observations, will com- colours) distribution, of comparable stel- Scognamiglio, D. et al. 2020, ApJ, 893, 4
Spiniello, C. et al. 2021a, A&A, 646, A28
prise 1D spectra for all the 52 objects, of lar masses and with optical spectra avail- Spiniello, C. et al. 2021b, arXiv:2103.12086
which half will likely be confirmed as rel- able. In particular, for the normal-sized Szomoru, D., Marijn, F. & van Dokkum, P. G. 2012,
ics, based on the current results (10 out ETGs we will derive mass-weighted ages ApJ, 749, 121
of 19). With more objects at our disposal, and metallicities, and thus star formation Tortora, C. et al. 2016, MNRAS, 457, 2845
Tortora, C. et al. 2018, MNRAS, 481, 4728
we will further investigate possible corre- histories, using the same code and under Trujillo, I. et al. 2009, ApJL, 692, L118
lations between the degree of relicness similar assumptions and setups to those van Dokkum, P. G. et al. 2008, ApJL, 677, L5
and other galaxy properties (size, mass, used in S21b. This will allow us to assess
colours) and/or with the environments in whether relics are simply the ultra-compact
Links
which they sit. tail of the distribution of red and dead
ETGs or if they are special and very rare 1
The INSPIRE project webpage: https://sites.
The final INSPIRE catalogue of confirmed systems with a completely different evo- google.com/inaf.it/chiara-spiniello/inspire/
relics, all with precise structural, kine- lutionary path. 2
The Kilo Degree Survey: http://kids.strw.
leidenuniv.nl
matic and stellar population parameters 3
The ESO VST Public Surveys: https://www.eso.
measured, multi-band wide-field optical Thanks to the high quality of the spectra, org/sci/observing/PublicSurveys/sciencePublic
and NIR photometry and high signal-to- we will measure very precise (< 5% uncer- Surveys.html#VST
noise spectroscopic data, will represent tainties) integrated velocity dispersion 4
The INSPIRE DR1 ESO Phase 3 collection on the
an important benchmark for cosmological values (σ ) for each system. This is one Science Archive Portal: http://archive.eso.org/
* scienceportal/home?data_collection=INSPIRE
simulations that should reproduce the of the most fundamental tracers of the
number density and the morphological, stellar populations and halo masses of
dynamical and stellar characteristics galaxies, and hence can act as a valuable
of both relics and younger ultra-compact identifier of a consistent population over
galaxies. cosmic time. Measuring σ from spectro-
*
scopic data and performing dynamical
Since different models predict a different models, we will infer dynamical masses.
redshift evolution of the number density Then, estimating precise stellar masses
of relics, a statistically significant cata- via stellar population analysis, we will be

A spectacular lunar halo — known as a 22° halo —
Astronomical News formed in the sky above ESO’s La Silla Observa-

tory. The optical phenomenon is a result of moon-
light interacting with ice crystals suspended in the
atmosphere, forming a ring with an apparent
radius of approximately 22° around the moon. It is
also known as the “moon ring” or “winter halo”.
ESO/Y. Beletsky (LCO)
Astronomical News DOI: 10.18727/0722-6691/5242
Maintaining Scientific Discourse During a

Global Pandemic: ESO’s First e-Conference #H02020
Richard I. Anderson 1, 2, 3
Sherry H. Suyu 4, 5, 6 Zoom Live stream
YouTube
Registration required Registration optional
Antoine Mérand 1
Moderators
1
ESO Speakers Live
2
ORIGINS cluster, University Observatory participants
Munich, Germany
3
Laboratory of Astrophysics, Institute of Panelists Live Q&A
Physics, EPFL, Lausanne, Switzerland on Slido
4
Max Planck Institute for Astrophysics,
Garching, Germany
5 Discussion panel
Technical University of Munich,
Garching, Germany
6
Academia Sinica Institute of Astronomy
and Astrophysics, Taipei, Taiwan
Live and asynchronous discussion on Slack (registered participants)
From 22 to 26 June 2020, ESO hosted
its first live e-conference, #H02020, from Figure 1. The #H02020 online meeting setup.
within its Headquarters in Garching,
Germany. Every day, between 200 and astrometry to theoretical cosmology. Participant surveys before & after the
320 participants around the globe tuned The conference was planned to run from meeting
in to discuss the nature and implications Monday to Friday, featuring 20 invited
of the discord between precise determi- talks, 2 introductory lectures, 34 contrib- Recognising the experimental nature
nations of the Universe’s expansion rate, uted talks, 8 short 10-minute discussion of our e-meeting, we conducted partici-
H0. Originally planned as an in-person sessions, four 10-minute poster flash talk pant surveys before and after the meeting
meeting, we moved to the virtual domain sessions in addition to physical posters, to assess how well it met participant
to maintain strong scientific discourse and a 1.5-hour discussion panel at the needs and expectations, and to measure
despite the COVID-19 pandemic. Here end. A public lecture by Adam Riess at whether it fulfilled our own goals. Before
we describe our conference setup, feed- the Technical University of Munich’s the c
onference we collected responses
back gathered from participants before downtown campus was also planned. from 89 individuals, and 79 afterwards.
and after the meeting, and lessons Since anonymous submissions were
learned from this unexpected exercise. Registration for the workshop opened allowed, only 46 before & after responses
As e-conferences will become increas- in late January 2020 with abstracts due could be uniquely matched. The follow-
ingly common in the future, we provide by 1 April. However, as cases of COVID-19 ing account is based on this survey
our perspective on how they can make started skyrocketing in February and information.
scientific exchanges more effective March, it became clear that an in-person
and inclusive, and also climate friendly. meeting in June was unrealistic. Faced
with the question of whether to postpone Goals of the online meeting
or cancel the meeting, on 26 March we
Before 18 March 2020: in-person con- decided on a third option: to convert the In adapting the conference for the virtual
ference at ESO Headquarters, Garching in-person meeting to a live global e-con- domain, we pursued the following main
ference within the course of 12 weeks. goals:
Our preparations for an in-person confer-
ence involving approximately 100 partici- We had no blueprint to follow and no pre- 1. To advance the specific research field
pants began in summer 2019, shortly after arranged technical solutions for implement- by enabling continued international
the Kavli Institute for Theoretical Physics ing ESO’s first global live e-conference, scientific exchange.
workshop on Tensions between the Early leading to a somewhat tricky situation that 2. To create a schedule compatible with
and the Late Universe (Verde, Treu & was rendered even less straightforward most regions of the world.
Riess, 2019). Together with the Scientific by the fact that nearly all ESO staff were 3. To foster informal discussions that
Organising Committee (SOC), composed working remotely for the first time. None- go beyond the scope of the invited
of Chuck Bennett, Annalisa Calamida, theless, we felt strongly that maintaining presentations.
Matthew Colless, Frédéric Courbin, strong international scientific discourse 4. To strike a balance between giving
Claudia de Rham, Wolfgang Gieren, Chow- was worth the challenge — especially the wider community a strong voice
Choong Ngeow, Hiranya Peiris, Mickael since many other meetings were being and covering a broad range of topics
Rigault, Dan Scolnic and Licia Verde, we cancelled or postponed indefinitely — and via invited presentations.
selected invited speakers who would we decided to go ahead with this meet- 5. To explore and leverage benefits
cover a wide range of topics from basic ing despite the short planning timescale. inherent to online meetings, such

Astronomical News Anderson, R. I. et al., ESO’s First e-Conference #H02020
as reduced access barriers to foster Figure 2. Word cloud

created from questions
diversity, equity, and inclusion.
on Slido. 301 questions
6. To provide inspiration and guidance were asked, 48% of
for those considering hosting an them anonymously, and
e-conference. they received a total of
1168 upvotes.
The meeting setup was chosen so as to

maximise the above goals, to the possible
detriment of other worthwhile goals, such
as networking sessions or online poster
sessions focusing on young researchers,
which we unfortunately could not include
given the compactified planning schedule.
Meeting setup
Figure 1 illustrates the e-conference setup

adopted, which was also explained in
a YouTube video1. Targeting goal number in the Zoom call in order to ensure an Live participation on Zoom and YouTube
2 above, we selected the time slot of orderly and uncomplicated meeting. To (up to 330 live participants on Day 1)
12:50–15:10 UTC on each day, which is at add a human touch, each day’s session exceeded our in-person expectations
least somewhat close to normal waking began and ended with a brief greeting three-fold, while asynchronous streaming
hours in most areas of the world. However, during which all Zoom participants were from YouTube has reached an audience
the start time (5:50 am in Vancouver) asked to turn on their video and greet 10 times greater than a conventional
and end time (1:10 am in Canberra) were each other. To provide a lively response to meeting would have done.
somewhat uncomfortable for regions each talk, moderators thanked speakers
bordering the Pacific Ocean. No breaks by applauding. This worked well enough, All questions were submitted via the
were included in the short window of but another possibility, suggested by Tom online platform Slido, which allowed par-
2 hours and 20 minutes. Feedback indi- Shanks, would be to use pre-recorded ticipants to upvote relevant questions
cated overall approval of this time slot. applause; despite being a bit artificial, it and supported anonymous question sub-
Some participants regretted the absence would provide something of the atmos- mission. Figure 2 shows a word cloud
of breaks, while others thought it appro- phere of a real conference and would be created from the questions submitted.
priate to skip breaks since most were fol- easier to manage than actual applause Questions deemed relevant by a majority
lowing the meeting from home. over many microphones on a conference were then relayed to speakers by the
call. session moderator. Advantages of Slido
Each day’s session consisted of four included very concise formulations as a
invited talks (20 minutes talk + 5 minutes Planning and adhering to a tight schedule result of the 160-character limit, democ-
Q&A) and a 30-minute live discussion were crucial to ensuring a smooth meet- ratisation of the Q&A thanks to voting,
panel 2. The latter served the dual purpose ing involving live participants across 18 the ability to ask questions anonymously,
of providing a voice to the community time zones. We therefore did extensive and a written account of all questions,
(in lieu of contributed talks) and enabling onboarding work with invited speakers which allowed speakers to reply in writing
critical discussion, thus targeting goals 3 and provided a short (< 5 minutes) YouTube to any unanswered questions. A simul
and 4 above. The panels featured pre- video explaining the meeting setup to all taneous benefit and drawback of using
pared statements and informal group dis- participants. Onboarding was done in Slido for Q&A was that follow-up ques-
cussion, and addressed questions left one-on-one Zoom calls during which we tions were not possible; the advantage
unanswered after the invited talks. Panels explained the meeting setup, launched being that precious Q&A time could not
were composed of the day’s invited mock presentations, and answered any be exhausted by individual questions
speakers and 3–5 participants who had technical questions. We believe that this and the disadvantage being that some
asked to be panel members during onboarding work was the reason so few questions were not answered satisfacto-
registration. technical issues were experienced during rily. Post-conference feedback shows
the conference. Although we briefly con- that a large majority of participants were
The conference call was held on Zoom sidered pre-recording talks to minimise extremely pleased with this form of Q&A.
and live streamed via YouTube 3 (incurring technical issues, we decided in favour of
a delay of ~ 20 s), where all videos remain live presentations to preserve the more We created a Slack workspace 4 for this
publicly accessible and have been viewed direct feeling of a real conference and to event to support asynchronous exchanges.
more than 5000 times. Only invited speak- avoid speakers’ feeling they had to record In principle, Slack allows the exchange
ers, discussion panelists, and moderators “the perfect talk”. of information in chat channels consisting
were invited to unmute their microphones of various user groups as well as allowing

participants to get in touch with each Italy | 5.4% Chile | 3.4%
other directly via one-to-one chat or video Portugal | 3.0%
Spain | 6.1% Iran | 2.7%
calls. Over the course of the conference Sweden | 2.7%
week, participants exchanged more than United Kingdom | 8.1% Mexico | 2.4%
1500 messages, including organisational Switzerland | 2.4%
messages, scientific discussions, answers Australia | 1.7%
India | 1.7%
to questions, diagrams and other docu- Poland | 1.4%
ments, such as presenter slides. How- Belgium | 1.0%
ever, becoming familiar with several new Canada | 1.0%
communication tools at once (Zoom, France | 12.8% Israel | 1.0%
Denmark | 0.7%
Slido, Slack) represented an initial hurdle Bulgaria | 0.7%
for some participants. Once accustomed, Colombia | 0.7%
though, participants reacted very posi- Argentina | 0.7%
tively to Slack as a coherent communica- Russia | 0.7%
Japan | 0.7%
tion platform for the meeting. The Netherlands | 0.7%
Turkey | 0.3%
Ireland | 0.3%
Participation and representation United States | 17.9% Brazil | 0.3%
Norway | 0.3%
Peru | 0.3%
We sought fair community representation Germany | 18.6% Serbia | 0.3%
among SOC members, who helped select
a diverse set of invited speakers across Figure 3. Registration locations by IP address.
the spectrum of career stages, locations,
gender, and subject matter. Registration
was open until the 300 person limit of observed trend suggests that women are website 5 and advertised it through ESO’s
the Zoom license was exhausted. The list less likely than men to continue beyond mailing list, newsletter and website, and
of participant host countries used to their doctorate degrees towards a career several astronomy-related mailing lists,
create Figure 3 was determined from par- in astronomy. However, we do not have also advertising it on Twitter. We are con-
ticipants’ IP addresses during registration. sufficient information to assess this drop- cerned that our announcement practices
The only information collected concern- off in detail, or to cross-reference it with could have excluded people from certain
ing the identity of participants was name, other potentially relevant factors, such regions, and we have identified that inter-
position, and affiliation. Gender statistics as postdoc-seniority or childcare respon- net censorship would have prevented
shown in Figure 4 are derived from manu- sibilities, which has disproportionately potential participants from China or Iran
ally (and possibly incorrectly) assigned impacted women during the COVID-19 from seeing our Twitter posts or YouTube
(binary) gender based on first names. pandemic. live streams. Professional organisations
such as the International Astronomical
Representation of women amongst the Representation of people of colour Union could play an important role in pro-
invited speakers and the SOC was signi amongst speakers and panelists was viding guidance on the best ways to
ficantly higher than amongst all regis- low, particularly in the case of non-Asian increase diversity in virtual conferences.
tered participants. Gender representation people of colour. This was especially
amongst panel members mirrored the evident given the timing of the meeting, Up to 120 participants (registered or not)
registration demographics. Two partici- which coincided with #BlackinSTEM week. joined via the YouTube livestream every
pants who had requested panel slots were There are many and complex reasons day. YouTube channel analytics provide
not assigned because they either misun- for the underrepresentation of people of information concerning audience age,
derstood the panel setup or because they colour in STEM, and they may differ sig- location, and gender, as shown in Fig-
had not yet completed graduate studies. nificantly across different regions of the ure 5. The analytics available are far from
Six of the 20 invited speakers were early- world. However, online conferences in complete since they are likely based on
career researchers. The all-male ‘other’ particular should strive to do better since inhomogeneous samples including regis-
category in Figure 4 comprises six retirees, barriers to participation (funding, travel, tered and unregistered YouTube users,
one teacher, one journalist, one freelance etc.) are lower than for in-person confer- depending on category. For example, the
researcher, and three other types that did ences. We note that we could not identify total number of views by location adds up
not fit any other category. any participants from the African conti- to only 46%; location information is miss-
nent based on either IP address, e-mail ing for the majority of viewers. Nonethe-
We note the sharp decrease in female address, host institute name, or YouTube less, we see that YouTube was favoured
representation between the categories analytics data, although we did have par- by a younger audience, possibly down to
of students and postdocs. It is a well- ticipants from several underprivileged different approaches to using the internet
known fact that retaining qualified women countries. We had listed our conference and/or different career stages.
is a major issue in STEM fields, and the on the Canadian Astronomy Data Centre

Female Male % views % watch time

100% 90
90%
80% 80
70%
60% 70
50%
40% 60
30% 50
20%
10% 40
0%
30
ts
cs
rs
ts
rs
rs
ed
r
f
he
af
lis
en
so
ke
be
do
er
st
ot
ne
ea
ud
es
em
st
st
64
20
pa
12
of
gi
sp
po
st
m
re
pr
87
22
20
62
C
5
70
SO
29
10
12
Figure 4. Career and gender demographics. Gender 0

was assigned according to first names provided. One
N Au ia
er pe
e
So Am lia
Am a
e
es – 34
ce
Eu a
es 44
es
4
es –24
K
Fr y
al
h eric
al
ic
an
–5
As
U
th tra
postdoc's name could not be assigned a gender.
an
i tl
ro
Ag 3 5 –
M
m
er
Ag 25
m
Ag 18
45
bt
s
Fe
Su
es
G
Ag
ut
or
Figure 5. YouTube channel analytics.
Goal 5: reducing access barriers to participant emphasised that these bene- A frequently mentioned drawback of the
increase diversity, equity, and inclusion fits far outweighed the inconvenience of virtual format is that it is less conducive
late-night sessions. to informal discussions than in-person
We deliberately collected no registration meetings. However, survey feedback also
fees in order to allow any interested par- Anonymous questions were perceived by revealed interesting potential benefits to
ties to participate. For comparison, the many as an important advantage over those discussions that did take place on
in-person registration fee would have classical Q&A. With no need to fear Slack. First, discussions on Slack were
been 180 euros (80 euros for students). embarrassment, many basic questions transparent to all participants instead
After converting our conference to an were submitted and voted for, in particu- of only a small group of people (for exam-
online format, we set up a new registra- lar during more theory-heavy talks. Irrele- ple, coffee-break clusters), helping non-
tion form. The registration served several vant questions — a concern for those specialists gain deeper insights “behind
purposes, including limiting access to opposed to anonymous questions — the scenes”. Second, discussions could
the Zoom call and Slack workspace, col- were not an issue because they were later be synthesised from the recorded
lecting consent to being recorded/live very few and did not get upvoted. The chat text. Third, one participant men-
streamed, and ensuring that participants 160-character limit of Slido was consid- tioned they felt more at ease entering into
pledged to abide by the ESO code of ered a challenge by some participants, an online discussion with strangers than
conduct. However, the entire conference although lengthier questions could also they would have done in person.
was publicly accessible even without be relayed to the Slack workspace.
registration via the YouTube live stream
and questions on Slido. All software tools A notable benefit of YouTube was the Overcoming shortcomings
used (Zoom, Slido, Slack) were free of ability to enable closed captioning (Eng-
charge to participants and compatible with lish subtitles). The average view duration The possibility for young scientists to
a maximum number of operating systems was 38 minutes 22 seconds with subti- present themselves and their work is cru-
with no installations required apart from tles enabled, compared to 20 minutes 59 cial to fostering their career development
a WebRTC capable browser (for example, seconds without subtitles; the 7.6% of and to providing a forum for “hot-off-the-
Firefox, Chrome, Opera). views with subtitles enabled accounted press” results that may revolutionise a
for 13% of the time watched on YouTube. field in the future. Unfortunately, in this
Most feedback mentioned that removing This underlines the need to assist partici- instance we were not able to arrange for
the need to travel was a key advantage pants in engaging with the materials pre- contributed talks by, for example, stu-
of the virtual format, irrespective of the sented, especially in the case of persons dents and early career scientists. How-
COVID-19 pandemic. Specific reasons with impaired hearing or non-native Eng- ever, feedback offered two very attractive
given included childcare, teaching and lish speakers. George Jacoby (NSF’s options for including early-career contri-
other work-related duties, care of pets, NOIRLab) further pointed out that partici- butions despite a tight live schedule: a)
visa-related issues, time saved by not pation conditions were more equal here pre-recorded contributed talks available
travelling, avoiding exhaustion (for exam- than in usual meetings because everyone for asynchronous viewing ahead of invited
ple, jet lag), health concerns, dietary sat in the front row with an unobstructed talks and discussion sessions; and b)
considerations, and more. An Australian view and adjustable speaker volume. online poster sessions on Slack. To pre-

vent exclusion by internet censorship, all erate the adoption of e-conferencing. “Are you now more likely to host e-conference?”
essential content should be rendered Now that the initial step has been taken,
accessible on platforms freely accessible e-conferences are becoming common-
worldwide. place and will likely become an integral
part of scientific discourse, not least
While e-conferences are not subject to because they are cheaper and more con-
travel-related access barriers, several venient for participants. Targeting our
other barriers may apply to the same goal number 6 above, we now close with
groups who would have difficulty attend- some final thoughts and recommendations.
ing in-person meetings. These include
internet censorship, lack of broadband First and foremost, feedback clearly
internet access, and the need for access shows that participant satisfaction with
to suitable personal devices (for example, this conference (on both the scientific and
1 laptop per person), among other technical level) was high, and it is worth
aspects that should be considered even noting that no-one stated that the confer-
before planning virtual meetings. One ence had been worse than expected. No Yes Maybe
way of addressing such issues could be Instead, many were positively surprised,
to create regional viewing hubs once and a majority stated that this meeting “How effective did you find Q&A on Slido?”
health measures allow it, as outlined by increased the likelihood of their organis-
Reshef et al. (2020). ing an e-conference themselves. The
Q&A sessions on Slido in particular were
Perhaps the most common negative a highlight for most, with 58% (46/79)
feedback from participants was that they saying the Q&A was very effective, 25%
had wished for more opportunities for somewhat effective, and only 1/79 indi-
discussion. The key to improving this cating that it was somewhat ineffective.
aspect of e-conferencing seems to be to Several people even stated they hoped
motivate participants to commit to offline for Slido-style Q&A to be incorporated in
discussions. Obstacles to this include a future face-to-face meetings.
lack of engagement (intentional or circum-
stantial) and the complexity of orienting Of course, valid criticisms exist about
oneself in a Slack workspace. During e-conferences and their limitations, nota-
#H02020, we witnessed participants bly concerning the lack of direct human
become increasingly engaged as they interaction. Participants particularly noted Neutral Not much
learned how to use this tool. In particular the following issues, in arbitrary order: Effective Highly effective
the ability of direct video calls amongst a) the lack of body language; b) difficulty
Slack workspace members seems to meeting new people; c) missing off-the-
have been underused. Instructions for record discussions; d) the missing cul- effort of attending conferences from the
how to use Slack should therefore be tural elements of international travel that professional to the private domain, for
part of the onboarding information, and are crucial to fostering understanding example regarding food & drink, physical
using the same tools in all conference- across cultures, languages, etc.; e) miss- (office) space, computing resources, etc.
related communications would lower the ing networking opportunities; and f) miss-
need for participants to familiarise them- ing opportunities for senior and junior At the same time, classical in-person
selves with new tools each time. Addition- researchers to meet. Some of these conferences also have significant short-
ally, asking participants to specify key- drawbacks that may seem unsurmount comings that are easily forgotten or over-
words upon registration could help to able now may be addressed by smart looked, perhaps because the community
assign discussion groups, connect partici- e-conference design or may evolve as is used to them. Classical conferences
pants according to interests, and foster social norms evolve, for example regard- are biased both implicitly (for example,
networking. ing (online) networking etiquette. How- stereotypes, personalities) and explicitly
ever, the human aspect of conferences (for example, ability to travel), resulting
is vital to scientific exchange and must in unintended exclusionary practices or
Final thoughts and recommendations not be neglected. situations. As a result, in-person meetings
prioritise specific types of human interac-
The fact that e-conferencing is much more E-conferences may also harbour long- tions over other interactions that could
climate-friendly than classical conferences term negative side effects. For example, lead to other benefits, such as transparent
is widely known in the community and privacy concerns remain underdiscussed, discussions, and foster equity, diversity
the carbon savings of e-conferences have despite recorded live streams’ drastically and inclusion.
been described in detail by, for example, changing the dynamic and persistence
Jahnke et al. (2020). Nonetheless, it took of participant contributions. Additionally, Social media are the native communica-
a worldwide health emergency to accel- e-meetings tend to shift the costs and tion platforms for e-conferences. Organ-

isers should include social media in their key aspect of scientific exchange. Com- the Technical University of Munich for providing
the venue for the public talk that was originally
announcement strategies and seek advice bining the two options would benefit from
planned for the in-person conference, and Karin
from press or outreach offices to make the best of both worlds, and we believe Lichtnecker and Gudrun Obst for the arrangements.
adequate use of social media, bearing in this could be achieved by increasing R. I. Anderson acknowledges support through an
mind the possibility of censorship. coordination amongst conference organ- ESO Fellowship, the ORIGINS excellence cluster
v isitor programme, and from the Swiss National
ising committees. For example, one could
Science Foundation through an Eccellenza Profes-
Another issue to solve is that of the com- imagine a few very large international sorial Fellowship (award No. 194638). S. H. Suyu
plexity of conference software. At the meetings per year that primarily focus on thanks the Max Planck Society for support through
time, our review of the available software networking and personal interactions, the Max Planck Research Group, and the European
Research Council (ERC) for support under the Euro-
revealed no single solution that would once health issues allow. These in-person
pean Union’s Horizon 2020 research and innovation
have incorporated all the functionality networking events could be complemented program (LENSNOVA: grant agreement No 771776).
desired for our meeting, so we adopted a by frequent e-conferences focused on
hybrid approach involving Zoom/YouTube, presenting scientific results, with the pos-
References
Slack, and Slido. Although corporate sibility for networking events centered
solutions do exist, they can be costly and around regional screenings. At the same Jahnke, K. et al. 2020, Nature Astronomy, 4, 812
were thus not applicable for an event with time, free online seminar series, such as Reshef, O. et al. 2020, Nature Reviews Materials,
zero registration fees. A detailed compar- the Golden Webinar series8 or the ESO 5, 253
Verde, L., Treu, T. & Riess, A. G. 2019, Nature
ison between different software solutions Cosmic Duologues9, are offering new
Astronomy, 3, 891
is beyond the scope of this article, but opportunities for scientific exchange.
should certainly be conducted. The docu-
mentation and research gathered by the The scientific community has entered a Links
EuroPython conference series6 offer new era of possibilities for scientific 1
Video explaining the e-conference setup: https://
substantial insights concerning the tech- exchange. We argue that there could be www.youtube.com/watch?v=j2HkdQpc7fc&t=1s
nical aspects of online conferencing. immense benefits on the horizon, and 2
Webpage with details of the H02020 conference:
reducing our carbon footprint is certainly https://www.eso.org/sci/meetings/2020/H0.html
3
It is not clear that this conference could one of them. However, the drawbacks, Conference YouTube page: https://www.youtube.
com/channel/UCs0UolX-xdD_hU8Rl-qSbLA
have taken place perhaps just 5 years challenges, and potential dangers of 4
Slack workspace: https://h02020.slack.com
ago. Projecting this rapid progress into e-conferencing should not simply be 5
Canadian Astronomy Data Centre website:
the future, new tools and solutions will ignored. A larger conversation should https://www.cadc-ccda.hia-iha.nrc-cnrc.gc.ca/en/
become available that may lead to even therefore consider how e-conferencing meetings/
6
more options to choose from, and thus will become a safe, inclusive, and carbon- EuroPython Society discussion of online events:
https://www.europython-society.org/
more complexity for conference attend- friendly addition to the landscape of inter- post/617463429296472064/sharing-our-research-
ees. To facilitate virtual-domain scientific national scientific discourse. and-licenses-for-going-online
cooperation in the future, it may be useful 7
Mattermost open source collaboration platform:
to define or develop a platform or set of https://mattermost.com
8
tools (perhaps an open source version Acknowledgements Golden Webinars: https://www.uc.cl/en/news/
golden-webinars-in-astrophysics
of Slack, such as CERN’s Mattermost 7, We thank all the participants of #H02020, and espe- 9
ESO Cosmic Duologues: https://www.eso.org/sci/
hosted by an institute) that the commu- cially the speakers, panelists, SOC members, and meetings/garching/Cosmic-Duologues.html
nity can adopt over a time span and that the many participants who provided feedback. 10
Conference reddit site: https://www.reddit.com/r/
renders the learning curve worthwhile. cosmology/comments/hdwjgp/eso_conference_
We thank Stella Chasiotis-Klingner and Nelma Alas h0_2020_assessing_uncertainties_in/
Whilst the creative and competitive devel- Da Cunha Dias Da Silva for their administrative
opment of new tools certainly leads to assistance, ESO safety and logistics personnel for
innovation, standardisation would make it helping us set up the H02020 control centre in a safe
worthwhile for scientists to invest time in environment during a health pandemic, and Marius
Chelu for setting up the room infrastructure.
becoming familiar with such channels.
Many colleagues showed enthusiasm for an
We believe that e-conferencing can e-conference and contributed ideas that made this
replace a majority of in-person confer- meeting a success. These include (in no particular
order): Jason Spyromilio, Michael Hilker, Giacomo
ences in the future, and we see many Beccari, Remco van der Burg, Marianne Heida,
good reasons for this development. These Dominika Wylezalek, Anna Miotello, Carlo Manara,
notably include environmental benefits, Sara Mancino, ESO Director for Science Rob Ivison,
time savings, and opportunities for ESO Director General Xavier Barcons, and many
others. We also thank Behnam Javanmardi for
increasing diversity, equity and inclusion, c overing the conference on reddit 10. We thank the
as well as the ability to create new con- Max Planck Institute for Astrophysics (MPA) for
ference formats that allow improved p roviding access to their Zoom license, and Andreas
global scientific discourse. Yet the human Weiss at MPA for IT support. We thank Slack and
Slido for offering free trial versions of their platforms,
element of scientific exchange must also which made this conference setup possible, and
be satisfied for science to progress. YouTube for offering the free live streaming capability
Hence, in-person meetings will remain a as well as the ability to revisit videos later. We thank

Astronomical News DOI: 10.18727/0722-6691/5243
Fellows at ESO
Paulo A. Miles-Páez started to study physics at the Universi-

dad Complutense de Madrid (UCM). The
Astronomy has captivated my attention (old) Licentiate in Physics degree con-
since I was a child, just like many of my sisted of five years of study. Three of gen-
colleagues. I was born in La Serena eral physics, in which I got a taste of
(Chile), a town almost unknown to most research work by doing a couple of sum-
of the world, but very popular amongst mer projects in optics and solid state
astronomers. The first time I saw a dark, physics, and two final years of specialisa-
deep sky was when I was four or five tion — when obviously I chose astro-
years old. It was during a visit to my great- physics. Thanks to the Bologna Process,
grandparents, who lived in a small village my generation had to take an extra aca-
in the mountains near the Argentinian demic year to get the new MSc degrees
border. I still remember the darkness of before qualifying to start a PhD pro-
that sky and the large collection of stars gramme in Spain. This final year was
that populated it, some of them so faint mostly research-focused, so I investi-
that I was not sure if they were real or gated the activity properties of ultra-cool
just my eyes tricking me. After this, one dwarfs in the infrared under the super
of my hobbies was to watch the Moon vision of David Montes — as preparatory
and some areas of the sky with a pair of work for the Calar Alto high-Resolution
binoculars from my backyard every night. search for M dwarfs with Exoearths with
Local outreach activities also contributed Near-infrared and optical Échelle Spec-
to my passion for the night sky. During the trographs (CARMENES) instrument. Apart
mid-1990s a group of astronomers used from the study time, I also have very
to visit public schools in La Serena with a good memories of my days in the faculty lots of hours using robotic telescopes
portable planetarium. I am not sure which playing cards with my colleagues — more from home.
astronomical association organised this than 10 years later, when we all are in
project, but I am very grateful as this was Madrid we still meet each other to have After defending my PhD, I moved to the
the first time that somebody talked to some drinks and play again. University of Western Ontario, Canada,
me about the things that are going on in to start my first postdoc in Stan Metchev’s
the sky. This fascinated me so much that By the end of my degree, I had developed substellar objects team. There I continued
I used to ask my parents to take me to a particular interest in brown dwarfs and my studies on substellar atmospheres
other schools just to visit the planetarium extrasolar planets, so I decided to start by using photometric and spectroscopic
again. a PhD. This was possible in 2011 when I observations with data from Gemini,
was offered the Astrofísico Residente Hubble, Spitzer, and some robotic tele-
Unfortunately, in the 1990s doing astron- fellowship at the Instituto de Astrofísica scopes from the Southeastern Associa-
omy in La Serena was considered an de Canarias (IAC) in Tenerife. Leaving my tion for Research in Astronomy (SARA)
eccentric activity, mostly reserved to the comfort zone in Madrid was not an easy network. I also had the opportunity to
American and European researchers who decision, but I was resolved to perform observe at Kitt Peak National Observa-
visited the town before heading to the my research at an observatory. For my tory in Arizona (another privileged place
observatories in the area. When I started PhD I worked with Enric Pallé and María to contemplate the night sky), and helped
to ask how to become an astronomer, the Rosa Zapatero Osorio on the atmos- in the initial construction of the Colibri
best option that I found was to move to pheric characterisation of very low-mass Telescope Array near London, Ontario,
Santiago and try there. In addition, I was stars and brown dwarfs using observa- sometimes in winter at –20 degrees C!
always advised (at school!) to study some- tions of linear polarisation. At the IAC I had I am very thankful to my wife, who was
thing “more useful” and “practical” in order the opportunity to interact with research- crazy enough to join me on this adventure
to secure a “good job”. In this regard, my ers interested in topics as diverse as solar in North America, where our son was
parents and my little sister were crucial, physics, stellar evolution and cosmology born. By the end of my postdoc in 2018,
as they always supported me and gave (and everything in between). I also got a I was offered an ESO Fellowship in Munich.
me the freedom to choose any path for lot of observing experience at the obser- ESO is one of the key names that I heard
my career. vatories of Izaña (Tenerife) and Roque de as a child, and working there was a
los Muchachos (La Palma). Some of my chance that I could not miss, so my little
In 2001 my quiet life in La Serena changed favourite telescopes during my PhD were family and I switched continents again in
completely, as my family moved to Madrid the IAC80, the Nordic Optical Telescope September 2019.
(Spain). I was so surprised (and confused) and the William Herschel Telescope (and
the first time I saw the Orion constellation obviously the FOcal Reducer and low In my research I try to understand the
upside-down! In the following years I dispersion Spectrograph 2 [FORS2] on atmospheric structure and composition
could not see very dark skies in the capi- the Very Large Telescope [VLT], too!). At of substellar objects by using different
tal of Spain, but I did visit the planetarium the end of my thesis I had accumulated techniques and instruments from several
in Madrid many times. After high school I about 100 nights of observations, and telescopes. This has made me realise the

Astronomical News Miles-Páez, P. A. & Gentile Fusillo, N. P., Fellows at ESO
importance of the (sometimes invisible) was 16 I won a scholarship to attend the

tasks that take place in an observatory — United World College of the Atlantic, a
from the moment that a proposal is boarding school in Wales, so I moved from
approved to the time when data are col- Italy to the UK. There I was very lucky to
lected. Because of this I decided to do meet teachers who were passionate
my functional work at the ESO User Sup- about their subjects and eager to foster
port Department. I currently give support the same passion in their students. I will
to programmes related to FORS2 and the forever be grateful to my physics teacher,
Echelle SPectrograph for Rocky Exoplanet Gabor Vincze, who recognised my interest
and Stable Spectroscopic Observations in astronomy and helped me decide to
(ESPRESSO), and soon to the CRyogenic apply to study physics at university.
high-resolution InfraRed Echelle Spectro-
graph+ (CRIRES+), too. I enjoy checking I stayed in the UK and enrolled in the
the observational strategies that other integrated Master’s physics course at
astronomers send us for their programmes, the University of Warwick where for my
and helping them when they get stuck in Master’s I “mined” the database of
the elaboration of this material. My duties rejected objects from the Wide Angle
give me the chance to interact with our Search for Planets (WASP) looking for
colleagues from Paranal, and to learn from variable stars. This was my first real
them when they identify an observing experience of a research environment in
run that can be optimised to maximise the astronomy, and, despite its unique mix
science return. I am also involved in the of frustration and excitement, it convinced
preparatory work for the upgrade of me that I wanted to take my education
FORS (FORS-Up), which allows me to further and apply for a PhD. I was offered solar planetary systems. In my research I
learn about the initial phases in the the opportunity to stay at the University had the opportunity to use the ESO tele-
design of an instrument. Apart from my of Warwick and pursue a PhD with Boris scopes at Paranal Observatory in Chile
duties, I am currently the Fellow mentor Gänsicke on the topic of white dwarfs. and naturally became very interested in
of four PhD students at ESO, and have Warwick became my second home and I the international organisation behind
participated in the organisation of the dedicated myself to absorbing as much these cutting-edge facilities. With a sci-
second and third ESO Summer Pro- as I could from the amazing people who ence focus in observational astronomy
grammes; these have been extremely made up the astronomy group. I will never and my previous experience of large tele-
rewarding experiences. be able to thank Boris enough for all that scopes, I was soon drawn to the ESO
he taught me as my PhD supervisor. Fellowship. I joined ESO in Garching as
During my free time I enjoy exploring the a fellow in 2019 and immediately felt
beautiful area of Bavaria with my family, During my doctorate, I also took part in a welcome in the diverse social and scien-
that recently got bigger with the birth of one-year studentship in the Isaac Newton tific environment of this organisation. I
my daughter, or visiting some very old Group on the island of La Palma, where I had originally hoped that my functional
buildings. Astronomy has set out the path worked as a support astronomer at the work would be to support operations
that I have followed for my career, taking Isaac Newton Telescope (INT). The INT is at Paranal, but after just my first trip to
me to several countries, living lots of a remarkable facility where most of the Chile the global pandemic made this
adventures and meeting people from time one lone astronomer spends night duty option impossible. However, at ESO
different cultures. Every minute invested after night running the 2.5-metre tele- there is no lack of opportunities to take
in learning about it has been worth it! scope completely on their own. Those part in extremely interesting projects
were strange, but also somewhat magical, and in my new functional work I use my
nights and the INT proved to be the per- expertise in white dwarfs and large area
Nicola Pietro Gentile Fusillo fect “school” for observational astronomy. surveys to find standard stars for the
I treasure every single day of my year in next-generation instruments that will be
I was six years old when for the first time I La Palma and the experience renewed mounted on the ELT.
looked through a telescope and observed my passion for astronomy. Back at
the marks that the comet Shoemaker- Warwick I completed my PhD and almost Despite the chaotic year we have just
Levy had left on Jupiter. I couldn’t know it immediately started a postdoc position been through, ESO remains a unique
at the time, but the fascination I felt then, in the same group under the leadership place where I have found absolute free-
realising I was directly seeing events of Pier-Emmanuel Tremblay. I was given dom in my science objectives, an amaz-
happening on another world, would never considerable freedom to pursue my own ing and supportive group of peers, and
leave me. research interest and focused much of an open international environment where
my work on constructing large samples in the space of one coffee break problems
When I was growing up astronomy of white dwarfs using the Gaia satellite and doubts are answered with expertise
became one of my passions and it drove and then using these white dwarfs as and wild ideas can turn into successful
my academic interest in physics. When I tools to explore the evolution of extra- projects.

Astronomical News
Personnel Movements
Arrivals (1 July 2021– 30 September 2021) Departures (1 July 2021– 30 September 2021)
Europe Europe
Amiaux, Jérôme (FR) Instrumentation Engineer/Physicist Beuchert, Tobias (DE) ESO Supernova Presenter
Claes, Rik (BE) Student Cerpa Urra, Nelly Natalia (CL) Student
Langan, Ivanna (FR) Student Davison, Thomas (UK) Student
Mandla, Christopher (DE) Electronics Engineer Erkal, Jessica (IE) Student
Reiter, Megan (US) Fellow Facchini, Stefano (IT) Fellow
Somigliana, Alice (IT) Student Fahrion, Katja (DE) Student
Thomson-Paressant, Keegan (FR/AU) Student Gil, Virginie (FR) Administrative Assistant
Tutuntzisz, Tamas (HU) IT Specialist – Infrastructure Pala, Anna Francesca (IT) Fellow
van Marrewijk, Joshiwa (NL) Student Petit dit de la Roche, Dominique (NL) Student
Weng, Simon (AU) Student Rombout, Francky (BE) Payroll Officer
Sánchez Menguiano, Laura (ES) Fellow
Stanke, Thomas (DE) User Support Astronomer
Wallace, Jane (UK) Council Secretary
Chile Chile
Beauchesne, Benjamin (FR) Student Ahumada, Bernardo (CL) Mechanical Section Supervisor
Frensch, Yolanda (NL) Student De Figueiredo Melo, Claudio (BR) ESO Representative in Chile
Behara, Natalie (FR) Data Scientist Gran, Felipe (CL) Student
Labdon, Aaron (UK) Fellow Houllé, Mathis Gilles Vital (FR) Student
Kakkad, Darshan (IN) Fellow
Sedaghati, Elyar (IR) Fellow
ALMA (ESO/NAOJ/NRAO)/Benisty et al.
This image, taken with the Atacama Large Millimeter/submillimeter

Array (ALMA), shows wide (left) and close-up (right) views of the
moon-forming disc surrounding PDS 70c, a young Jupiter-like
planet nearly 400 light-years away. The close-up view shows
PDS 70c and its circumplanetary disc centre-front, with the larger
circumstellar ring-like disc taking up most of the right-hand side of
the image. The star PDS 70 is at the centre of the wide-view image
on the left (for more details see ESO press release eso2111).

The Messenger 184

Uploaded by

Document Informationclick to expand document informationThe Messenger is a quarterly journal presenting ESO's activities to the public.

Document Informationclick to expand document information

Copyright:

Available Formats

The Messenger 184

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Messenger 184

Uploaded by

Copyright:

Available Formats

Impact of Satellite Constellations and ESO’s Role in Supporting the Community

2020 Scientific Prioritisation Community Poll

The Messenger is published, in hardcopy

Printed by omb2 Print GmbH,

Unless otherwise indicated, all images in

2 The Messenger 184 | 2021

Analysing the Impact of Satellite Constellations and

The Messenger 184 | 2021 3

Observatory latitude: – 24.6° Constellation: Observatory latitude: – 24.6° Constellation:

Number of satellites degree – 2

4 The Messenger 184 | 2021

10 – 2 the satellite from direct illumination, and

ments of the attitude of the satellites have

Imager also reduced their brightness: the orien-

Low-res spectrograph tation of the solar panel has been adjusted

The Messenger 184 | 2021 5

6 The Messenger 184 | 2021

Percentage brightness reduction

The Messenger 184 | 2021 7

Report on the Scientific Prioritisation Community Poll (2020)

5 years 2 years or less

My home institution is located in: Details for “Other”: Figure 1. (Above)

8 The Messenger 184 | 2021

than 80% of the respondents are from Ultraviolet/

The Messenger 184 | 2021 9

10 The Messenger 184 | 2021

Wide-field and/or low-angular-resolution

0 5 10 15 20 0.3 to 0.4 to 1 to 3 to Sub-mm Radio 0.3 to 0.4 to 1 to 3 to Sub-mm Radio

The Messenger 184 | 2021 11

Other 2.2% Other 8.0%

Data Archive 58.6% SPICA (Space Infrared

12 The Messenger 184 | 2021

Fast turnaround observing programmes Figure 9. (Middle) The relevance to respondents’

The Messenger 184 | 2021 13

Structure and evolution of

Milky Way dynamics

Life cycle of stars

Search for life

The Sun and the

Time domain astronomy

Large mm/sub-mm single-

dish (30–50 m in diameter)

Very long baselines

ACA (mm/sub-mm compact

APEX (single-dish sub-mm)

VLTI (1.8-m or 8-m telescope

ELT (40-m-class optical/IR

VLT (8-m-class optical/IR

14 The Messenger 184 | 2021

This lone antenna is part of the Atacama Large

in full display: green airglow hovers above the hori-

A Guide to ALMA Operations and Interactions with

16 The Messenger 184 | 2021

Amongst the numerous subsystems,

The Messenger 184 | 2021 17

18 The Messenger 184 | 2021

Shutdown sational structure, its functioning as an

Cioni, M. et al. 2019, The Messenger, 176, 8