Serb. Astron. J. } 176 (2008), 1 - 13 UDC 520.

2
DOI: 10.2298/SAJ0876001I Invited review

LARGE SYNOPTIC SURVEY TELESCOPE:

FROM SCIENCE DRIVERS TO REFERENCE DESIGN

Ž. Ivezić1 , T. Axelrod2 , W. N. Brandt3 , D. L. Burke4 , C. F. Claver5 , A. Connolly1 ,

K. H. Cook6 , P. Gee7 , D. K. Gilmore4 , S. H. Jacoby2 , R. L. Jones1 , S. M. Kahn4 ,
J. P. Kantor2 , V. Krabbendam5 , R. H. Lupton8 , D. G. Monet9 , P. A. Pinto10 , A. Saha5 ,
T. L. Schalk11 , D. P. Schneider3 , M. A. Strauss7 , C. W. Stubbs12 , D. Sweeney2 ,
A. Szalay13 , J. J. Thaler14 , and J. A. Tyson7 for the LSST Collaboration

1
Department of Astronomy, University of Washington, Box 351580, Seattle, WA 98195
E–mail: [email protected]

2
LSST Corporation, 4703 E. Camp Lowell Drive, Suite 253, Tucson, AZ 85712
3
Department of Astronomy and Astrophysics, The Pennsylvania State University,
525 Davey Lab, University Park, PA 16802
4
Kavli Institute for Particle Astrophyics and Cosmology, Stanford Linear Accelerator Center,
Stanford University, Stanford, CA, 94309
5
National Optical Astronomy Observatory, 950 N. Cherry Ave, Tucson, AZ 85719
6
Lawrence Livermore National Laboratory, 7000 East Avenue, Livermore, CA 94550
7
Physics Department, University of California, One Shields Avenue, Davis, CA 95616
8
Department of Astrophysical Sciences, Princeton University, Princeton, NJ 08544
9
U.S. Naval Observatory Flagstaff Station, 10391 Naval Observatory Road, Flagstaff, AZ 86001
10
Steward Observatory, The University of Arizona, 933 N Cherry Ave., Tucson, AZ 85721
11
University of California–Santa Cruz, 1156 High St., Santa Cruz, CA 95060
12
Departments of Physics and Astronomy, Center for Astrophysics, Harvard University,
60 Garden St., Cambridge, MA 02138
13
Department of Physics and Astronomy, The John Hopkins University,
3701 San Martin Drive, Baltimore, MD 21218
14
University of Illinois, Physics and Astronomy Departments,1110 W. Green St., Urbana, IL 61801

(Received: March 5, 2008; Accepted: March 5, 2008)

1
 Ž. IVEZIĆ et al.

SUMMARY: In the history of astronomy, major advances in our understanding

of the Universe have come from dramatic improvements in our ability to accurately
measure astronomical quantities. Aided by rapid progress in information technology,
current sky surveys are changing the way we view and study the Universe. Next-
generation surveys will maintain this revolutionary progress. We focus here on the
most ambitious survey currently planned in the visible band, the Large Synoptic
Survey Telescope (LSST). LSST will have unique survey capability in the faint time
domain. The LSST design is driven by four main science themes: constraining dark
energy and dark matter, taking an inventory of the Solar System, exploring the
transient optical sky, and mapping the Milky Way. It will be a large, wide-field
ground-based system designed to obtain multiple images covering the sky that is
visible from Cerro Pachón in Northern Chile. The current baseline design, with
an 8.4 m (6.5 m effective) primary mirror, a 9.6 deg2 field of view, and a 3,200
Megapixel camera, will allow about 10,000 square degrees of sky to be covered using
pairs of 15-second exposures in two photometric bands every three nights on average.
The system is designed to yield high image quality, as well as superb astrometric and
photometric accuracy. The survey area will include 30,000 deg2 with δ < +34.5◦ ,
and will be imaged multiple times in six bands, ugrizy , covering the wavelength
range 320–1050 nm. About 90% of the observing time will be devoted to a deep-
wide-fast survey mode which will observe a 20,000 deg2 region about 1000 times in
the six bands during the anticipated 10 years of operation. These data will result
in databases including 10 billion galaxies and a similar number of stars, and will
serve the majority of science programs. The remaining 10% of the observing time
will be allocated to special programs such as Very Deep and Very Fast time domain
surveys. We describe how the LSST science drivers led to these choices of system
parameters.

Key words. Astronomical data bases: miscellaneous – Atlases – Catalogs – Surveys

– Solar system: general – Stars: general – Galaxy: general – Galaxies: general –
Cosmology: miscellaneous

1. INTRODUCTION dertake large-scale sky surveys. As vividly demon-

strated by surveys such as the Sloan Digital Sky
Survey (SDSS, York et al. 2000), the Two Micron
1.1. Large scale surveys: a new way of seeing All Sky Survey (2MASS, Skrutskie et al. 2006), and
the Galaxy Evolution Explorer (GALEX, Martin et
Major advances in our understanding of the al. 2006), to name but a few, sensitive and accu-
Universe have historically arisen from dramatic im- rate multi-color surveys over a large fraction of the
provements in our ability to ”see”. We have de- sky enable an extremely broad range of new scien-
veloped progressively larger telescopes over the past tific investigations. These results, based on synergy
century, allowing us to peer farther into space, and of advances in telescope construction, detectors, and
further back in time. With the development of above all, information technology, have dramatically
advanced instrumentation – imaging, spectroscopic, impacted nearly all fields of astronomy – and many
and polarimetric – we have been able to parse radia- areas of fundamental physics. In addition, the re-
tion detected from distant sources over the full elec- cent world-wide attention received by Google Sky1
tromagnetic spectrum in increasingly subtle ways. (Scranton et al. 2007) demonstrates that the impact
These data have provided the detailed information of sky surveys extends far beyond fundamental sci-
needed to construct physical models of planets, stars, ence progress and reaches all of society.
galaxies, quasars, and larger structures. Motivated by the evident scientific progress
Until recently, most astronomical investiga- made possible by large sky surveys, three recent
tions have focused on small samples of cosmic sources nationally endorsed reports by the U.S. National
or individual objects. This is because our largest
telescope facilities have rather small fields of view, Academy of Sciences2 concluded that a dedicated
typically only a few square arcminutes – a tiny frac- ground-based wide-field imaging telescope with an
tion (few parts per hundred million) of the sky, and effective aperture of 6–8 meters is a high priority for
those with large fields of view could not detect very planetary science, astronomy, and physics over the
faint sources. With all of our existing telescope fa- next decade. The Large Synoptic Survey Telescope
cilities, we have still surveyed only a minute volume (LSST) described here is such a system. The LSST
of the observable Universe. will be a large, wide-field ground based telescope de-
Over the past two decades, however, advances signed to obtain multi-band images over a substan-
in technology have made it possible to move beyond tial fraction of the sky every few nights. The survey
the traditional observational paradigm and to un- will yield contiguous overlapping imaging of over half

1 http://earth.google.com/sky/

2 Astronomy and Astrophysics in the New Millennium, NAS 2001; Connecting Quarks with the Cosmos: Eleven Science Ques-
tions for the New Century, NAS 2003; New Frontiers in the Solar System: An Integrated Exploration Strategy, NAS 2003.

2
 LSST: FROM SCIENCE DRIVERS TO REFERENCE DESIGN

the sky in six optical bands, with each sky location 2.1. The Main Science Drivers
visited about 1000 times over 10 years.
The purpose of this paper is to provide a sum- The main science drivers are used to optimize
mary of the main LSST science drivers and how they numerous system parameters. Ultimately, in this
led to the current system design parameters, as de- high-dimensional parameter space, there is a one-
scribed in §2. A project status report and concluding dimensional manifold defined by the total project
remarks are presented in §3. For detailed and up-to- cost. The science drivers must both justify this cost,
date information, please consult the LSST website as well as provide guidance on how to optimize var-
(www.lsst.org). ious parameters while staying on the cost manifold.
Here we summarize a dozen most important inter-
locking constraints on data properties imposed by
2. THE LSST REFERENCE DESIGN the four main science themes:
o The depth of a single visit (an observation con-
sisting of two back-to-back exposures of the
The most important characteristic that deter- same region of sky)
mines the speed at which a system can survey a given o Image quality
sky area to a given depth (faint flux limit) is its o Photometric Accuracy
étendue (or grasp), the product of its primary mir- o Astrometric Accuracy
o Optimal exposure time
ror area and the field-of-view area (assuming that o The filter complement
observing conditions such as seeing, sky brightness, o The distribution of revisit times (i.e. the ca-
etc., are fixed). The effective étendue for LSST will dence of observations)
be greater than 300 m2 deg2 , which is more than o The total number of visits to a given area of
an order of magnitude larger than that of any exist- the sky
ing facility. For example, the SDSS, with its 2.5-m o The coadded survey depth
telescope (Gunn et al. 2006) and a camera with 30 o The distribution of visits on the sky, and the
imaging CCDs (Gunn et al. 1998), has an effective total sky coverage
étendue of only 7.5 m2 deg2 . o The distribution of visits per filter
The range of scientific investigations which o Data processing and data access (e.g. time
will be enabled by such a dramatic improvement in delay for reporting transient sources and the
survey capability is extremely broad. Guided by the software contribution to measurement errors)
community-wide input assembled in the report of the
Science Working Group of the LSST3 , the LSST de- We present a detailed discussion of how these
sign is focused to achieve goals set by four main sci- science-driven data properties are transformed to
ence themes: system parameters below.

(1) Constraining Dark Energy and Dark Matter 2.2. Constraining Dark Energy and
Dark Matter
(2) Taking an Inventory of the Solar System
(3) Exploring the Transient Optical Sky Current models of cosmology require the exis-
(4) Mapping the Milky Way tence of both dark matter and dark energy to match
observational constraints (Spergel et al. 2007). Dark
Each of these four themes itself encompasses energy affects the cosmic history of both the Hub-
a variety of analyses, with varying sensitivity to in- ble expansion and mass clustering. If combined, dif-
strumental and system parameters. These themes ferent types of probes of the expansion history and
fully exercise the technical capabilities of the system, structure history can lead to percent level precision
such as photometric and astrometric accuracy and in dark energy and other cosmological parameters.
image quality. The working paradigm is that all sci- These tight constraints arise because each technique
entific investigations will utilize a common database depends on the cosmological parameters or errors in
constructed from an optimized observing program, different ways. These probes include weak gravita-
such as that discussed in Section 3. Here we briefly tional lens (WL) cosmic shear, baryon acoustic os-
describe the science goals and the most challenging cillations (BAO), supernovae, and cluster counting –
requirements for the telescope and instruments that all as a function of redshift. Using the cosmic mi-
are derived from those goals, which led to the overall crowave background as normalization, the combina-
system design decisions discussed below. For a more tion of these probes can yield the needed precision
detailed discussion, we refer the reader to the LSST to distinguish between models of dark energy (Zhan
Science Requirements Document4 , as well as to the 2006, and references therein). In addition, time-
numerous LSST poster presentations at the recent resolved strong galaxy and cluster lensing probes the
211th Meeting of the AAS5 . physics of dark matter. This is because the positions

3 Available as http://www.lsst.org/Science/docs/DRM2.pdf
4 Available at http://www.lsst.org/Science/docs.shtml
5 Available at http://www.lsst.org/Meetings/AAS/2008/AAS211.shtml

3
 Ž. IVEZIĆ et al.

and profiles of multiple images of a source galaxy bands will be obtained (due to the spread in red-
depend sensitively on the total mass distribution, in- shift). The importance of K-corrections to supernova
cluding the dark matter, in the lensing object. cosmology implies that the calibration of the relative
While LSST WL and BAO probes will yield offsets in photometric zero points between filters and
the strongest dark energy and dark matter con- the knowledge of the system response functions, es-
straints, two major programs from this science theme pecially near the edges of bandpasses, must be accu-
that provide unique and independent constraints on rate to about 1% (Wood-Vasey et al. 2007). Deeper
the system design are data (r > 26) for a small area of the sky can ex-
o Weak lensing of galaxies, and tend the discovery of SN to a mean redshift of 0.7,
o Type Ia Supernovae. with some objects beyond z ∼1. The added sta-
Weak lensing (WL) techniques can be used to tistical leverage on the ”pre-acceleration” era would
map the distribution of mass as a function of redshift improve constraints on the properties of dark energy
and thereby trace the history of both the expansion as a function of redshift.
of the universe and the growth of structure (e.g. Hu 2.3. Taking an Inventory of the Solar System
and Tegmark 1999, Wittman et al. 2000, for a review
see Bartelmann and Schneider 2001). These investi- The small-body populations in the Solar Sys-
gations require deep wide-area multi-color imaging tem, such as asteroids, trans-Neptunian objects
with stringent requirements on shear systematics in (TNOs) and comets, are remnants of its early as-
at least two bands, and excellent photometry in all sembly. The history of accretion, collisional grind-
bands. The strongest constraints on the LSST image ing, and perturbation by existing and vanished giant
quality come from this science program. In order to planets is preserved in the orbital elements and size
control systematic errors in shear measurement, it is distributions of those objects. In the main asteroid
mandatory to obtain the desired depth with many belt between Mars and Jupiter collisions still occur,
short exposures (which effectively enables ”random- and occasionally objects are ejected on orbits that
ization” of systematic errors). Detailed simulations may take them on a collision course with the Earth.
of weak lensing techniques show that, in order to As a result, the Earth orbits within a swarm
obtain a sample of ∼3 billion lensing galaxies, the of asteroids; some number of these objects will ul-
coadded map must cover ∼20,000 deg2 , and reach a timately strike Earth’s surface. In December 2005,
depth of r ∼ 27.5 (5σ for point sources), with several the U.S. Congress directed6 NASA to implement a
hundred exposures per field and sufficient signal-to- near-Earth object (NEO) survey that would catalog
noise in at least five other bands to obtain accurate 90% of NEOs larger than 140 meters by 2020. About
photometric redshifts (Zhan 2006). Because of their 20% of NEOs, the potentially hazardous asteroids or
low surface brightness, this depth optimizes the num- PHAs, are in orbits that pass sufficiently close to
ber of detected galaxies in ground-based seeing, and Earth’s orbit, to within 0.05 AU, that perturbations
allows their detection in significant numbers to be- with time scales of a century can lead to intersec-
yond a redshift of two. It is anticipated that opti- tions and the possibility of collision. In order to ful-
mal science analysis of weak lensing will place strong fill the Congressional mandate using a ground-based
constraints on data processing software, such as si- facility, a 10-meter class telescope equipped with a
multaneous analysis of all the available data (Tyson multi-gigapixel camera, and a sophisticated and ro-
et al. 2008). bust data processing system are required (Ivezić et
Type Ia supernovae (SN) provided the first ev- al. 2007). The search for NEOs also places strong
idence that the expansion of the universe is acceler- constraints on the cadence of observations, requiring
ating (Riess et al. 1998, Perlmutter et al. 1999). To closely spaced pairs of observations (two or prefer-
fully exploit the supernovae science potential, light- ably three times per lunation) in order to link obser-
curves sampled in multiple bands every few days over vations unambiguously and derive orbits. Individ-
the course of a few months are required. This is ual exposures should be shorter than about 1 minute
essential to search for systematic differences in su- each to minimize the effects of trailing for the major-
pernovae populations which may masquerade as cos- ity of moving objects. The images must be well sam-
mological effects, as well as to determine photomet- pled to enable accurate astrometry, with absolute ac-
ric redshifts from the supernovae themselves. Un- curacy of at least 0.1 arcsec. The images should reach
like other cosmological probes, even a single object a depth of at least ∼24.5 (5σ for point sources) in the
can provide useful constraints and, therefore, a large r band in order to probe the ∼ 0.1 km size range at
number of SN across the sky can enable a high an- main-belt distances, and to fulfill the Congressional
gular resolution search for any dependence of dark NEO mandate. The photometry should be better
energy properties on direction, which would be an than 1-2% to enable color-based taxonomic classifi-
indicator of new physics. cation.
Given the expected SN flux distribution, the
single visit depth should be at least r ∼ 24. Good 2.4. Exploring the Transient Optical Sky
image quality is required to separate SN photomet- Recent surveys have shown the power of vari-
rically from their host galaxies. Observations in at ability for studying gravitational lensing, searching
least five photometric bands are necessary to ensure for supernovae, determining the physical properties
that, for any given supernova, light-curves in several of gamma-ray burst sources, and many other projects

6 For details see http://neo.jpl.nasa.gov/neo/report2007.html

4
 LSST: FROM SCIENCE DRIVERS TO REFERENCE DESIGN

at the forefront of astrophysics (Tyson 2006, and ref- servation to enable parallax and proper motion mea-
erences therein). A wide-area dense temporal cov- surements; and dynamic range that allows measure-
erage to deep limiting magnitudes would enable the ments of astrometric standard stars at least as bright
discovery and analysis of rare and exotic objects such as r = 16. In order to probe the halo out to its
as neutron star and black hole binaries, gamma-ray presumed edge at ∼ 100 kpc using numerous main-
bursts and X-ray flashes, at least some of which ap- sequence stars, the total co-added depth must reach
parently mark the deaths of massive stars; AGNs r > 27, with a similar depth in the g band. To study
and blazars; and very possibly new classes of tran- the metallicity distribution of stars in the Sgr tidal
sients, such as binary mergers and stellar disruptions stream (see e.g. Majewski et al. 2003) and other halo
by black holes. It is likely that such a survey would substructures at distances beyond the presumed in-
detect numerous microlensing events in the Local ner vs. outer halo boundary (at least ∼ 40 kpc), the
Group and perhaps beyond, and open the possibility co-added depth in the u band must reach ∼ 24.5.
of discovering planets and obtaining spectra of lensed To detect RR Lyrae stars beyond the Galaxy’s tidal
stars in distant galaxies as well as our own. radius at ∼ 300 kpc, the single-visit depth must be
Time domain science requires large area cov- r ∼ 24.5. In order to constrain the tangential ve-
erage to enhance the probability of detecting rare locity of stars at a distance of 10 kpc, where halo
events; good time sampling, since light curves are dominates over disk, to within 10 km/s needed to
necessary to distinguish certain types of variables be competitive with large-scale radial velocity sur-
and in some cases to infer their properties (e.g. de- veys, the required proper motion accuracy is at least
termining of the intrinsic luminosity of supernovae 0.2 mas/yr. The same accuracy follows from the
Type Ia depends on measurements of their rate of de- requirement to obtain the same proper motion ac-
cline); accurate color information to assist with the curacy as Gaia (Perryman et al. 2001) at its faint
classification of variable objects; good image quality limit (r ∼ 20). In order to produce a complete sam-
to enable discerning of images, especially in crowded ple of solar neighborhood stars out to a distance of
fields; and rapid data reduction, classification and 300 pc (the thin disk scale height), with 3σ or better
reporting to the community in order to flag interest- geometric distances, trigonometric parallax measure-
ing objects for spectroscopic and other investigations ments accurate to 1 mas are required. To achieve the
with separate facilities. Time scales ranging from 1 required proper motion and parallax accuracy with
min (to constrain the properties of fast faint tran- an assumed astrometric accuracy of 10 milliarcsec
sients such as optical flashes associated with gamma- per observation per coordinate, approximately 1,000
ray bursts (Kaspi et al. 2007) and transients recently observations are required. This requirement on the
discovered by the Deep Lens Survey, Becker et al. number of observations is in good agreement with
2004) to 10 years (to study long-period variables and the independent constraint implied by the difference
quasars) should be probed over a significant fraction between the total depth and the single-visit depth.
of the sky. It should be possible to measure colors of
fast transients, and to reach faint magnitude limits 2.6. A Summary and Synthesis of
in individual visits (at least the Deep Lens Survey Science-driven
limit of r ∼ 24.5). Constraints on Data Properties
2.5. Mapping the Milky Way The goals of all the science programs dis-
A major objective of modern astrophysics is cussed above (and many others, of course) can be ac-
to understand when and how galaxies formed and complished by satisfying the following minimal con-
evolved. Theories of galaxy formation and evolution straints7
can be tested and influenced by a significantly im- o The single visit depth should reach r ∼ 24.5.
proved understanding of the distribution and kine- This limit is primarily driven by the NEO
matics of stars in our own Galaxy, the Milky Way,
which is a complex and dynamical structure that is survey and variable sources (e.g. RR Lyrae
still being shaped by the infall (merging) of neigh- stars), and by proper motion and trigono-
boring smaller galaxies. We still lack robust answers metric parallax measurements for stars. Indi-
to two basic questions about the Milky Way Galaxy: rectly, it is also driven by the requirements on
the coadded survey depth and the minimum
o What is the detailed structure and accretion number of exposures placed by weak lensing
history of the Milky Way? science.
o What are the fundamental properties of all the o Image quality should maintain the limit set
stars within 300 pc of the Sun? by the atmosphere (the median seeing is 0.7
arcsec in the r band at the chosen site), and
Key requirements for mapping the Galaxy are not be degraded appreciably by the hardware.
large area coverage, excellent image quality to max- In addition to stringent constraints from weak
imize the photometric and astrometric accuracy, es- lensing, good image quality is driven by re-
pecially in crowded fields; photometric precision of quired survey depth for point sources and by
at least 1% to separate main sequence and giant image differencing techniques.
stars; astrometric precision of about 10 mas per ob-

7 Fora more elaborate listing of various constraints, including detailed specification of various probability distributions, please
see the LSST Science Requirements Document (http://www.lsst.org/Science/docs.shtml).

5
 Ž. IVEZIĆ et al.

o Photometric repeatability should achieve 5 mil- bandpasses (because the system throughput
limag precision at the bright end, with zero- and atmospheric properties are wavelength de-
point stability across the sky of 10 millimag pendent, the achieved depths are different in
and band-to-band calibration errors not larger different bands). The adopted time allocation
than 5 millimag. These requirements are (see Table 1) gives a slight preference to the
driven by the photometric redshift accuracy, r and i bands because of their dominant role
the separation of stellar populations, detec- for star/galaxy separation and weak lensing
tion of low-amplitude variable objects (such as measurements.
eclipsing planetary systems), and the search o The distribution of visits on the sky should ex-
for systematic effects in type Ia supernova tend over at least ∼20,000 deg2 to obtain the
light-curves. required number of galaxies for weak lensing
o Astrometric precision should maintain the studies, with attention paid to ”special” re-
limit set by the atmosphere, of about 10 mil- gions such as the Ecliptic, Galactic plane, and
liarcsec per visit at the bright end (on scales the Large and Small Magellanic Clouds.
below 20 arcmin). This precision is driven by o Data processing, data products and data access
the desire to achieve a proper motion accuracy should enable efficient science analysis with-
of 0.2 mas/yr and parallax accuracy of 1.0 mas out a significant impact on the final uncertain-
over the course of a 10 year long survey. ties. To enable a fast and efficient response
o The single visit exposure time should be less to transient sources, the processing latency
than about a minute to prevent trailing of fast should be less than a minute, with a robust
moving objects and to facilitate control of var- and accurate preliminary classification of re-
ious systematic effects induced by the atmo- ported transients.
sphere. It should be longer than ∼ 20 seconds It is remarkable that, even with these joint re-
to avoid efficiency losses due to finite readout quirements, none of the individual science programs
and slew time. is severely overdesigned. That is, despite their sig-
o The filter complement should include at least nificant scientific diversity, these programs are highly
six filters in the wavelength range limited by compatible in terms of desired data characteristics.
atmospheric absorption and silicon detection Indeed, any one of the four main science drivers could
efficiency (320–1050 nm), with roughly rectan- be removed, and the remaining three would still yield
gular filters and no large gaps in the coverage, very similar requirements for most system parame-
in order to enable robust and accurate photo- ters. As a result, the LSST system can adopt a highly
metric redshifts, and stellar typing. An SDSS- efficient survey strategy where a single dataset serves
like u band is extremely important for sepa- all science programs (instead of science-specific sur-
rating low-redshift quasars from stars, and for veys executed in series). One can think of this as
estimating metallicity of F/G main sequence massively parallel astrophysics. The vast majority
stars. A bandpass with an effective wave- (about 90%) of the observing time will be devoted
length of about 1 micron would enable studies to a deep-wide-fast survey mode, with the remaining
of sub-stellar objects, high-redshift quasars,
and regions of the Galaxy that are obscured 10% of observing time allocated to special programs
by interstellar dust. which will also address multiple science goals. Before
o The revisit time distribution should allow SN describing these surveys in more detail, we discuss
light curves to be sampled every few days; this the main system parameters.
constraint is needed to obtain orbits of Solar 2.7. The Main System Design Parameters
System objects as well, while accomodating
constraints set by proper motion and trigono- Given the minimum science-driven constraints
metric parallax measurements. on the data properties listed in the previous section,
o The total number of visits of any given area we now discuss how they are translated into con-
of sky, when accounting for all filters, should straints on the main system design parameters: the
be of the order of 1,000, as mandated by weak aperture size, the optimal exposure time, and the fil-
lensing science, the NEO survey, and proper ter complement. We also briefly describe the LSST
motion and trigonometric parallax measure- reference design.
ments. Studies of transient sources also bene- 2.8. The Aperture Size
fit from a larger number of visits.
o The coadded survey depth should reach r ∼ The product of the system’s étendue and the
27.5, with sufficient signal-to-noise ratio in survey lifetime, for given observing conditions, deter-
other bands to address both extragalactic and mines the sky area that can be surveyed to a given
Galactic science drivers. depth, where the étendue is the product of the pri-
o The distribution of visits per filter should mary mirror area and the field-of-view area. The
enable accurate photometric redshifts, sepa- LSST field-of-view area is maximized to its practical
ration of stellar populations, and sufficient limit, 10 deg2 , determined by the requirement that
depth to make detection of faint extremely the delivered image quality be dominated by atmo-
red sources posible (e.g. brown dwarfs and spheric seeing at the chosen site (Cerro Pachón in
high-redshift quasars). Detailed simulations Northern Chile). A larger field-of-view would lead
of photometric redshift estimates suggest an to unacceptable deterioration of the image quality.
approximately flat distribution of visits among This leaves the primary mirror diameter and survey

6
 LSST: FROM SCIENCE DRIVERS TO REFERENCE DESIGN

lifetime as free parameters. The adopted survey life- 2.9. The Optimal Exposure Time
time of 10 years is a compromise between a shorter
time that leads to an excessively large and expensive The single visit depth depends on both the pri-
mirror (15 m for a 3 year-long survey and 12 m for mary mirror diameter and the chosen exposure time.
In turn, the exposure time determines the time in-
a 5-year long survey), and a smaller telescope that terval to revisit a given sky position and the total
would require more time to complete the survey, with number of visits, and each of these quantities has
the associated increase in operations cost. its own science drivers. We summarize these simul-
The primary mirror size is a function of the taneous constraints in terms of single-visit exposure
required survey depth and the desired sky coverage. time:
By and large, the anticipated science outcome scales
with the number of detected sources. For practically o The single-visit exposure time should not be
all astronomical source populations, in order to max- longer than about a minute to prevent trail-
imize the number of detected sources, it is more ad- ing of fast Solar System moving objects, and
vantageous to maximize the area first, and then the to enable efficient control of atmospheric sys-
detection depth. For this reason, the sky area for tematics.
the main survey is also maximized to its practical o The mean revisit time (assuming uniform ca-
limit, 20,000 deg2 , determined by the requirement dence) for a given position on the sky, n
to avoid large airmasses (which would substaintially (days), scales as
deteriorate the image quality and the survey depth).
With the adopted field-of-view area, the sky µ
texp
¶µ
Asky
¶µ
10 deg2
¶
coverage and the survey lifetime fixed, the primary n= ,
mirror diameter is fully driven by the required survey 10 sec 20, 000 deg2 AFOV
depth. There are two depth requirements: the final (1)
(coadded) survey depth, r ∼ 27.5, and the depth of a where the losses for realistic observing condi-
single visit, r ∼ 24.5. The two requirements are com- tions have been taken into account. Science
patible if the number of visits is several hundred (per drivers such as SN and moving objects in the
Solar System require that n < 4, or equiva-
band), which is in good agreement with independent lently texp < 40 seconds for the nominal values
science-driven requirements on the latter. of Asky and AFOV . Note that normalization
The required coadded survey depth provides a
direct constraint, independent of the details of sur- by 20,000 deg2 is equivalent to two visits per
vey execution such as the exposure time per visit, on night over 10,000 deg2 .
the minimum primary mirror diameter, as illustrated o The number of visits to a given position on
in Fig. 1. the sky, Nvisit , with losses for realistic observ-
ing conditions taken into account, is given by
µ ¶µ ¶
3000 T
Nvisit = . (2)
n 10 yr

The requirement Nvisit > 800, again implies

that n < 4 and texp < 40 seconds if the survey
lifetime, T ∼ 10 years.
o These three requirements place a firm up-
per limit on the optimal exposure time of
texp < 40 seconds. Surveying efficiency (the
ratio of open-shutter time to the total time
spent per visit) considerations place a lower
limit on texp due to finite read-out and slew
time (the longest acceptable read-out time is
set to 2 seconds, and the slew and settle time
is set to 5 seconds, including the read-out time
for the second exposure in a visit):
µ ¶
texp
ǫ= . (3)
texp + 9 sec
Fig. 1. The co-added depth in the r band vs. aper-
ture and the survey lifetime (r ∼ V , where V is the To maintain efficiency losses below 30% (i.e. at least
Johnson visual magnitude). It is assumed that 22% below the limit set by the weather patterns), and
of the total observing time (corrected for weather and to minimize the read noise impact, texp > 20 sec-
other losses) is allocated for the r band, and that the onds. Taking these constraints simultaneously into
ratio of the surveyed sky area to the field-of-view area account, as summarized in Fig. 2, yielded the follow-
is 2,000. ing reference design:
7
 Ž. IVEZIĆ et al.

1. A primary mirror effective diameter of ∼ 6.5 tions (Lenz et al. 1998, Helmi et al. 2003), and pho-
m. With the adopted optical design, described tometric selection of quasars (Richards et al. 2002).
below, this effective diameter corresponds to The extension of the SDSS system to longer wave-
a geometrical diameter of ∼ 8 m. Motivated lengths (the y band at ∼ 1 micron) is driven by the
by characteristics of the existing equipment at increased effective redshift range achievable with the
the Steward Mirror Laboratory, which is cast- LSST due to deeper imaging, the desire to study sub-
ing the primary mirror, the adopted geomet- stellar objects, high-redshift quasars, regions of the
rical diameter is set to 8.4 m. Galaxy that are obscured by interstellar dust, and
2. A visit time of 30 seconds (using two 15 sec- the scientific opportunity offered by modern CCDs
ond exposures to efficiently reject cosmic rays; with high quantum efficiency in the near infrared.
ǫ = 77%).
3. A revisit time of 3 days on average per 10,000
deg2 of sky, with two visits per night.
To summarize, the chosen primary mirror di-
ameter is the minimum diameter that simultaneously
satisfies the depth (r ∼ 24.5 for single visit and
r ∼ 27.5 for coadded depth) and cadence (revisit
time of 3-4 days, with 30 seconds per visit) con-
straints described above.

Fig. 3. The current design of the LSST band-

passes. The vertical axis shows the overall system
throughput. The computation includes the atmo-
spheric transmission, optics, and the detector sen-
sitivity.

2.11. The LSST Reference Design

We briefly describe the reference design for the

main LSST system components. Detailed discussion
of the flow-down from science requirements to sys-
tem design parameters, and extensive system engi-
neering analysis can be found in Claver et al. (2008,
in prep.). Additional discussion of science drivers,
description of data products and examples of sci-
ence programs can be found in Ivezić et al. (2008,
Fig. 2. The single-visit depth in the r band (5σ in prep.). Both publications will be maintained at
detection for point sources) vs. revisit time, n (or
the astro-ph site8 , and should be consulted for the
exposure time, texp = 10 n seconds), as a function detailed and most up-to-date information about the
of aperture size. In addition to direct constraints on LSST system.
optimal exposure time, texp is also driven by require-
ments on the revisit time, n, the total number of vis- 2.12. Telescope and Site
its per sky position over the survey lifetime, Nvisit , The large LSST étendue is achieved in a novel
and the survey efficiency, ǫ (see eqs.1-3). Note that three-mirror design (modified Paul-Baker, Davison
these constraints result in a fairly narrow range of and Angel 2002) with a very fast f/1.25 beam. The
allowed texp for the main deep-wide-fast survey.
optical design has been optimized to yield a large
2.10. The Filter Complement field of view (9.6 deg2 ), with seeing-limited image
quality, across a wide wavelength band (350–1050
The LSST filter complement (ugrizy, see
nm). Incident light is collected by the primary mir-
Fig. 3) is modeled after the Sloan Digital Sky Sur-
ror, which is an annulus with an outer diameter of
vey (SDSS) system (Fukugita et al. 1996) because of 8.4 m (an effective diameter of 6.5 m), then reflected
its demonstrated success in a wide variety of appli- to a 3.4 m convex secondary, onto a 5 m concave
cations, including photometric redshifts of galaxies tertiary, and finally into three refractive lenses in a
(Budavári et al. 2003), separation of stellar popula- camera (see Fig. 4). All three mirrors will be ac-
8 http://arxiv.org/archive/astro-ph

8
 LSST: FROM SCIENCE DRIVERS TO REFERENCE DESIGN

tively supported to control wavefront distortions in- 2.13. Camera

troduced by gravity and environmental stresses on The LSST camera provides a 3.2 Gigapixel
the telescope. flat focal plane array, tiled by 4K x 4K CCD sen-
sors with 10 µm pixels (see Figs. 5 and 6). This
pixel count is a direct consequence of sampling the
∼ 10 deg2 field-of-view with 0.2×0.2 arcsec2 pixels
(Nyquist sampling). The sensors are deep depleted,
back-illuminated devices with a highly segmented ar-
chitecture that enables the entire array to be read in
2 seconds.

Fig. 4. The LSST baseline optical design with its

unique monolithic mirror: the primary and tertiary
mirrors are so positioned that they form a continu-
ous compound surface, allowing them to be polished
into a single substrate.

The telescope mount is a compact, stiff struc-

ture with a fundamental frequency of nearly 10 Hz,
which is crucial for achieving the required fast slew-
and-settle times. The telescope sits on a concrete
pier within a carousel dome that is 30 m in diam-
eter. The dome has been designed to reduce dome
seeing (local air turbulence that can distort images)
and to maintain a uniform thermal environment over Fig. 6. The LSST focal plane. Each cyan square
the course of the night. The LSST Observatory will represents one 4096 × 4096 pixel large sensor. Nine
be sited atop Cerro Pachón in northern Chile, near sensors are assembled into a raft; the 21 rafts are
the Gemini South and SOAR telescopes (latitude: S outlined in red. There are 189 science sensors, each
30◦ 10’ 20.1”; longitude: W 70◦ 48’ 0.1”; elevation: with 16.8 Mpix, for a total pixel count of 3.2 Gpix.
2123 m; the median r band zenith seeing: 0.7 arcsec).
2.14. Data Management
The rapid cadence of the LSST observing pro-
gram will produce an enormous volume of data, ∼
30 TB per night, leading to a total database over the
ten years of operations of 60 PB for the raw data,
and 30 PB for the catalog database. The total data
volume after processing will be several hundred PB,
processed using substantial computing power (∼ 100
TFlops). Processing such a large volume of data,
converting the raw images into a faithful represen-
tation of the universe, and archiving the results in
useful form for a broad community of users is a ma-
jor challenge.
The data management system is configured in
Fig. 5. The LSST camera with a person to indicate three levels: an infrastructure layer consisting of the
scale size. The camera is positioned in the middle of computing, storage, and networking hardware and
the telescope and will include a filter mechanism and system software; a middleware layer, which handles
shuttering capability. distributed processing, data access, user interface,
and system operations services; and an applications
layer, which includes the data pipelines and products
and the science data archives. There will be both
mountain summit and base computing facilities, as
9
 Ž. IVEZIĆ et al.

Table 1. The LSST Baseline Design and Survey Parameters.

Quantity Baseline Design Specification
Optical/mount Configuration 3-mirror modified Paul-Baker; alt-azimuth
Final f-Ratio, aperture f/1.25, 8.4 m
Field of view area, étendue 9.6 deg2 , 318 m2 deg2
Plate Scale, pixel count 50.9 µm/arcsec (0.2” pix), 3.2 Gigapix
Wavelength Coverage, filters 320 – 1050 nm, ugrizy
Single visit depths (5σ) u : 23.9, g : 25.0, r : 24.7, i : 24.0, z : 23.3, y : 22.1
Mean number of visits u : 70, g : 100, r : 230, i : 230, z : 200, y : 200
Final (coadded) depths (5σ) u : 26.3, g : 27.5, r : 27.7, i : 27.0, z : 26.2, y : 24.9

well as a central archive facility and multiple data ac- visits for a ten-year LSST survey is 2,767,595 (∼ 5.5
cess centers. The data will be transported over exist- million 15-second long exposures). The per-band al-
ing high-speed optical fiber links from South America location of these visits is shown in Table 1. The re-
to the U.S. maining 10% of observing time will be used to obtain
improved coverage of parameter space such as very
2.15. The Baseline Main Deep-Wide-Fast deep (r ∼ 26) observations, observations with very
Survey
short revisit times (∼ 1 minute), and observations of
The fundamental basis of the LSST concept is ”special” regions such as the Ecliptic, Galactic plane,
to scan the sky deep, wide, and fast, and to obtain a and the Large and Small Magellanic Clouds.
dataset that simultaneously satisfies the majority of
science goals. This concept, so-called ”universal ca-
dence”, will yield the main deep-wide-fast survey and 3. CONCLUSIONS
use about 90% of the observing time. The observing
strategy will be optimized to maximize the scientific Until recently, most astronomical investiga-
throughput by minimizing slew and other downtime tions have focused on small samples of cosmic sources
and by making appropriate choices of the filter bands or individual objects. Over the past decade, however,
given the real-time weather conditions. As often as advances in technology have made it possible to move
possible, each field will be observed twice, with vis- beyond the traditional observational paradigm and
its separated by 15-60 minutes. This strategy will to undertake large-scale sky surveys, such as SDSS,
provide motion vectors to link detections of moving 2MASS, GALEX and many others. This observa-
objects in the Solar System, and fine-time sampling tional progress, based on synergy of advances in tele-
for measuring short-period variability. The resulting scope construction, detectors, and above all, infor-
sky coverage for LSST baseline cadence, based on mation technology, has a dramatic impact on nearly
detailed operations simulations, is shown for the r all fields of astronomy, many areas of fundamental
band in Fig. 7. The anticipated total number of physics, and the society in general.
The LSST builds on the experience of these
surveys and addresses the broad goals stated in sev-
eral nationally endorsed reports by the U.S. National
Academy of Sciences. The realization of the LSST
involves extraordinary engineering and technological
challenges: the fabrication of large, high-precision
optics; construction of a huge, highly-integrated ar-
ray of sensitive, wide-band imaging sensors; and the
operation of a massive data management facility han-
dling tens of terabytes of data each day. The project
is scheduled to have first light in 2014 and the begin-
ning of survey operations in 2015.
The LSST survey will open a movie-like win-
dow on objects that change brightness, or move, on
timescales ranging from 10 seconds to 10 years. The
Fig. 7. The distribution of the r band visits on the survey will have a data rate of about 30 TB/night
sky for the baseline main survey. The sky is shown (more than one complete Sloan Digital Sky Sur-
in Aitoff projection in equatorial coordinates and the vey per night), and will collect over 60 PB of raw
number of visits for a 10-year survey is color-coded data over its lifetime, resulting in an incredibly rich
according to the inset. The two regions with smaller and extensive public archive that will be a treasure
number of visits than the main survey are the Galac- trove for breakthroughs in many areas of astronomy.
tic plane (arc on the left) and the so-called ”northern About 10 billion galaxies and a similar number of
Ecliptic region” (upper right). It is likely that the re- stars will be detected – for the first time in history,
gion around the South Celestial Pole will also receive the number of cataloged celestial objects will exceed
substantial coverage. the number of living people!

10
 LSST: FROM SCIENCE DRIVERS TO REFERENCE DESIGN

LSST has been conceived as a public facility: REFERENCES

the database that it will produce, and the associ-
ated object catalogs that are generated from that
database, will be made available to the world’s sci- Bartelmann, M. and Schneider, P.: 2001, Physics
entific community and to the public at large with no Reports, 340, 291.
proprietary period. LSST will be a significant mile- Becker, A.C., Wittman, D.M., Boeshaar, P. C. et al.:
stone in the globalization of the information revolu- 2004, Astrophys. J., 611, 418.
tion. The LSST data management system will pro- Budavári, T., Connolly, A. J., Szalay, A. S. et al.:
vide user-friendly tools to access this database and 2003, Astrophys. J., 595, 59.
to support user-initiated queries, run on LSST com- Davison, W. and Angel, R.: 2002, in ”Survey and
puters, either at the archive facility or at the data Other Telescope Technologies and Discover-
access centers. We expect that many, perhaps even ies”, Tyson, J. A. and Wolff, S., eds. Pro-
the majority of LSST discoveries will come from re- ceedings of the SPIE, 4836, 110.
search astronomers with no formal affiliation to the Fukugita, M., Ichikawa, T., Gunn, J.E. et al.: 1996,
project, from students, and from interested ama- Astron. J., 111, 1748.
teurs, intrigued by the accessibility to the universe Gunn, J. E., Carr, M., Rockosi, C. et al.: 1998, As-
that this facility uniquely provides. tron. J., 116, 3040.
Gunn, J. E., Siegmund, W. A., Mannery, E. J. et al.:
2006, Astron. J., 131, 2332.
Acknowledgements – In 2003, the LSST Corporation Helmi, A., Ivezić, Ž., Prada, F. et al.: 2003, Astro-
was formed as a non-profit 501(c)3 Arizona corpora- phys. J., 586, 195.
tion with headquarters in Tucson, AZ. Membership Hu, W. and Tegmark, M.: 1999, Astrophys. J., 514,
has since expanded to more than twenty members L65.
Ivezić, Ž, Tyson, J. A., Jurić, M. et al.: 2007, Pro-
including Brookhaven National Laboratory, Califor- ceedings of IAU Symposium 236. Edited by
nia Institute of Technology, Carnegie Mellon Uni- G.B. Valsecchi and D. Vokrouhlický. Cam-
versity, Columbia University, Google Inc., Harvard- bridge: Cambridge University Press, 353 (also
Smithsonian Center for Astrophysics, Johns Hopkins
University, Kavli Institute for Particle Astrophysics astro-ph/0701506).
and Cosmology - Stanford University, Las Cum- Kaspi, S., Brandt, W. N., Maoz, D. et al.: 2007,
bres Observatory Global Telescope Network, Inc., Astrophys. J., 659, 997.
Lawrence Livermore National Laboratory, National Lenz, D. D., Newberg, J., Rosner, R. et al.: 1998,
Optical Astronomy Observatory, Princeton Univer- Astrophys. J. Suppl. Series, 119, 121.
sity, Purdue University, Research Corporation, Stan- Majewski, S. R., Skrutskie, M. F., Weinberg, M. D.
ford Linear Accelerator Center, The Pennsylvania and Ostheimer, J. C.: 2003, Astrophys. J.,
State University, The University of Arizona, Uni- 599, 1082.
versity of California at Davis, University of Cali- Martin, D. C., Fanson, J., Schiminovich, D. et al.:
fornia at Irvine, University of Illinois at Urbana- 2006, Astrophys. J., 619, L1.
Champaign, University of Pennsylvania, University Perlmutter, S., Aldering, G., Goldhaber, G. et al.:
of Pittsburgh, and the University of Washington. 1999, Astrophys. J., 517, 565.
LSST is a public-private partnership. Design and Perryman, M. A. C., de Boer, K. S., Gilmore, G. et
development activity is in part supported by the Na- al.: 2001, Astron. Astrophys., 369, 339.
tional Science Foundation under Scientific Program Richards, G. T., Fan, X., Newberg, H. J. et al.: 2002,
Astron. J., 123, 2945.
Order No. 9 (AST-0551161) and Scientific Program Riess, A. G., Filippenko, A. V., Challis, P. et al.:
Order No. 1 (AST-0244680) through Cooperative 1998, Astron. J., 116, 1009.
Agreement AST-0132798. Portions of this work are Scranton, R., Connolly, A., Krughoff, S. et al.: 2007,
supported by the Department of Energy under con- astro-ph/0709.0752
tract DE-AC02- 76SF00515 with the Stanford Linear Spergel, D. N., Bean, R., Doré, O. et al.: 2007, As-
Accelerator Center, contract DE-AC02- 98CH10886 trophys. J. Suppl. Series, 170, 377.
with Brookhaven National Laboratory, and contract Tyson, J. A.: 2006, Nature, 442, 364.
DE-AC52-07NA27344 with Lawrence Livermore Na- Tyson, J. A., Roat, C., Bosch, J. and Wittman,
tional Laboratory. Additional funding comes from D.: 2008, Astronomical Data Analysis Soft-
private gifts, grants to universities, and in-kind sup- ware and Systems XVII, ASP Conference Se-
port at Department of Energy laboratories and other ries, Vol. XXX, 2008; J. Lewis, R. Argyle, P.
LSSTC Institutional Members. NOAO is operated Bunclarck, D. Evans, and E. Gonzales-Solares,
by the Association of Universities for Research in As- eds.
tronomy, Inc. (AURA) under cooperative agreement Wittman, D. M., Tyson, J. A., Kirkman, D.,
with the National Science Foundation. KHC’s work Dell’Antonio, I. and Bernstein, G.: 2000, Na-
was performed under the auspices of the U.S. D.O.E. ture, 405, 143.
by LLNL under contract DE- AC52-07NA27344. Wood-Vasey, W. M., Miknaitis, G., Stubbs, C. W.
et al.: 2007, Astrophys. J., 666, 694.
York, D. G., Adelman, J., Anderson, S. et al.: 2000,
Astron. J., 120, 1579.
Zhan, H.: 2006, Journal of Cosmology and Astropar-
ticle Physics, 8, 8.

11
 Ž. IVEZIĆ et al.

LSST: OD NAUQNIH CIƨEVA DO DIZAJNA

Ž. Ivezić1 , T. Axelrod2 , W. N. Brandt3 , D. L. Burke4 , C. F. Claver5 , A. Connolly1 ,

K. H. Cook6 , P. Gee7 , D. K. Gilmore4 , S. H. Jacoby2 , R. L. Jones1 , S. M. Kahn4 ,
J. P. Kantor2 , V. Krabbendam5 , R. H. Lupton8 , D. G. Monet9 , P. A. Pinto10 , A. Saha5 ,
T. L. Schalk11 , D. P. Schneider3 , M. A. Strauss7 , C. W. Stubbs12 , D. Sweeney2 ,
A. Szalay13 , J. J. Thaler14 , and J. A. Tyson7 for the LSST Collaboration

1
Department of Astronomy, University of Washington, Box 351580, Seattle, WA 98195
E–mail: [email protected]

2
LSST Corporation, 4703 E. Camp Lowell Drive, Suite 253, Tucson, AZ 85712
3
Department of Astronomy and Astrophysics, The Pennsylvania State University,
525 Davey Lab, University Park, PA 16802
4
Kavli Institute for Particle Astrophyics and Cosmology, Stanford Linear Accelerator Center,
Stanford University, Stanford, CA, 94309
5
National Optical Astronomy Observatory, 950 N. Cherry Ave, Tucson, AZ 85719
6
Lawrence Livermore National Laboratory, 7000 East Avenue, Livermore, CA 94550
7
Physics Department, University of California, One Shields Avenue, Davis, CA 95616
8
Department of Astrophysical Sciences, Princeton University, Princeton, NJ 08544
9
U.S. Naval Observatory Flagstaff Station, 10391 Naval Observatory Road, Flagstaff, AZ 86001
10
Steward Observatory, The University of Arizona, 933 N Cherry Ave., Tucson, AZ 85721
11
University of California–Santa Cruz, 1156 High St., Santa Cruz, CA 95060
12
Departments of Physics and Astronomy, Center for Astrophysics, Harvard University,
60 Garden St., Cambridge, MA 02138
13
Department of Physics and Astronomy, The John Hopkins University,
3701 San Martin Drive, Baltimore, MD 21218
14
University of Illinois, Physics and Astronomy Departments,1110 W. Green St., Urbana, IL 61801

UDK 520.2
Pregledni rad po pozivu

U istoriji astronomije, veliki po- generacije nastavie ovim putem revolu-

maci u naxem razumevaƬu Vasione qesto cionarnog napretka. U ovom radu usredsre-
su proizlazili iz dramatiqnog napretka u ujemo se na najambiciozniji planirani pro-
mogunostima preciznog mereƬa astronom- jekat pregleda neba u vidƩivom delu spektra,
skih veliqina. ZahvaƩujui brzom razvoju Veliki sinoptiqki teleskop za pregled neba
informacionih tehnologija, savremeni pre- (skr. LSST, od eng. Large Synoptic Survey Tele-
gledi neba meƬaju naqin na koji posmatramo i scope). LSST e imati jedinstvene mogunosti
prouqavamo Vasionu. Pregledi neba sledee
12
 LSST: FROM SCIENCE DRIVERS TO REFERENCE DESIGN

pregleda u kratkim vremenskim intervali- trometrijsku i fotometrijsku taqnost. Pre-

ma. Dizajn LSST odreuju qetiri primarna gled e pokriti ukupnu povrxinu od 30 000
nauqna zadatka: ograniqavaƬe na parametre kvadratnih stepeni, u oblasti deklinacija
vezane za tamnu energiju i tamnu materiju, δ < +34.5◦ , snimajui vixe puta u xest fil-
pravƩeƬe inventara objekata Sunqevog sis- tera, ugrizy, koji pokrivaju oblasti talas-
tema, istraжivaƬe kratkotrajnih pojava na
nebu u vidƩivom delu spektra i mapiraƬe nih duжina od 320–1050 nm. Oko 90% posma-
Mleqnog puta. Teleskop e predstavƩati ve- traqkog vremena bie iskorixeno za rad u
liki, zemaƩski, xirokougaoni sistem dizajni- tzv. dubokom-xirokom-brzom modu, pri qemu
ran za dobijaƬe vixestrukih snimaka koji bi e se, tokom predvienih 10 godina rada
u potpunosti pokrili nebo vidƩivo iz mesta teleskopa, otprilike 1000 puta u xest fil-
Cerro Pachón u severnom Qileu. Aktuelni tera posmatrati oblast od 20 000 kvadrat-
osnovni dizajn predvia primarno ogledalo nih stepeni. PrikupƩeni podaci e biti poh-
preqnika 8.4 m (efektivno 6.5 m), vidno raƬeni u bazu koja e ukƩuqivati oko 10
poƩe od 9.6 kvadratnih stepeni i kameru sa milijardi galaksija i pribliжno isti broj
3200 megapiksela, xto e omoguiti da se u zvezda, i koja e sluжiti veini nauqnih pro-
dve ekspozicije od po 15 sekundi, u dva fo- grama. Preostalih 10% posmatraqkog vremena
tometrijska filtera, za tri noi u prose- predvieno je za posebne programe kao xto su
ku, pokrije ukupno 10 000 kvadratnih stepeni Vrlo Duboki i Vrlo Brzi pregledi. Ovde
opisujemo kako se od nauqnih zadataka pro-
neba. Sistem je dizajniran tako da obezbedi grama LSST doxlo do ovih izbora parametara
visok kvalitet snimaka, kao i izuzetnu as- sistema.

13