LSST: from Science Drivers to Reference Design and Anticipated Data Products
Željko Ivezić, Steven M. Kahn, J. Anthony Tyson, Bob Abel, Emily Acosta, Robyn Allsman, David Alonso, Yusra AlSayyad, Scott F. Anderson, John Andrew, James Roger P. Angel, George Z. Angeli, Reza Ansari, Pierre Antilogus, Constanza Araujo, Robert Armstrong, Kirk T. Arndt, Pierre Astier, Éric Aubourg, Nicole Auza, Tim S. Axelrod, Deborah J. Bard, Jeff D. Barr, Aurelian Barrau, James G. Bartlett, Amanda E. Bauer, Brian J. Bauman, Sylvain Baumont, Andrew C. Becker, Jacek Becla, Cristina Beldica, Steve Bellavia, Federica B. Bianco, Rahul Biswas, Guillaume Blanc, Jonathan Blazek, Roger D. Blandford, Josh S. Bloom, Joanne Bogart, Tim W. Bond, Anders W. Borgland, Kirk Borne, James F. Bosch, Dominique Boutigny, Craig A. Brackett, Andrew Bradshaw, William Nielsen Brandt, Michael E. Brown, James S. Bullock, Patricia Burchat, David L. Burke, Gianpietro Cagnoli, Daniel Calabrese, Shawn Callahan, Alice L. Callen, Srinivasan Chandrasekharan, Glenaver Charles-Emerson, Steve Chesley, Elliott C. Cheu, Hsin-Fang Chiang, James Chiang, Carol Chirino, Derek Chow, David R. Ciardi, Charles F. Claver, Johann Cohen-Tanugi, Joseph J. Cockrum, Rebecca Coles, Andrew J. Connolly, Kem H. Cook, Asantha Cooray, Kevin R. Covey, Chris Cribbs, Wei Cui, Roc Cutri, Philip N. Daly, Scott F. Daniel, Felipe Daruich, Guillaume Daubard, Greg Daues, William Dawson, Francisco Delgado, Alfred Dellapenna, Robert de Peyster, Miguel de Val-Borro, Seth W. Digel, Peter Doherty, Richard Dubois, Gregory P. Dubois-Felsmann, Josef Durech, Frossie Economou, Michael Eracleous, Henry Ferguson, Enrique Figueroa, Merlin Fisher-Levine, Warren Focke, Michael D. Foss, James Frank, Michael D. Freemon, Emmanuel Gangler, et al. (213 additional authors not shown)
DDraft version May 25, 2018
Typeset using L A TEX twocolumn style in AASTeX62
LSST: from Science Drivers to Reference Design and Anticipated Data Products ˇZeljko Ivezi´c, Steven M. Kahn,
2, 3
J. Anthony Tyson, Bob Abel, Emily Acosta, Robyn Allsman, David Alonso, Yusra AlSayyad, Scott F. Anderson, John Andrew, James Roger P. Angel, George Z. Angeli, Reza Ansari, Pierre Antilogus, Constanza Araujo, Robert Armstrong, Kirk T. Arndt, Pierre Astier, ´Eric Aubourg, Nicole Auza, Tim S. Axelrod, Deborah J. Bard, Jeff D. Barr, Aurelian Barrau, James G. Bartlett, Amanda E. Bauer, Brian J. Bauman, Sylvain Baumont,
16, 11
Andrew C. Becker, Jacek Becla, Cristina Beldica, Steve Bellavia, Federica B. Bianco,
19, 20
Rahul Biswas, Guillaume Blanc,
10, 22
Jonathan Blazek,
23, 24
Roger D. Blandford, Josh S. Bloom, Joanne Bogart, Tim W. Bond, Anders W. Borgland, Kirk Borne, James F. Bosch, Dominique Boutigny, Craig A. Brackett, Andrew Bradshaw, William Nielsen Brandt, Michael E. Brown, James S. Bullock, Patricia Burchat, David L. Burke, Gianpietro Cagnoli, Daniel Calabrese, Shawn Callahan, Alice L. Callen, Srinivasan Chandrasekharan, Glenaver Charles-Emerson, Steve Chesley, Elliott C. Cheu, Hsin-Fang Chiang, James Chiang, Carol Chirino, Derek Chow, David R. Ciardi, Charles F. Claver, Johann Cohen-Tanugi, Joseph J. Cockrum, Rebecca Coles, Andrew J. Connolly, Kem H. Cook, Asantha Cooray, Kevin R. Covey, Chris Cribbs, Wei Cui, Roc Cutri, Philip N. Daly, Scott F. Daniel, Felipe Daruich, Guillaume Daubard, Greg Daues, William Dawson, Francisco Delgado, Alfred Dellapenna, Robert de Peyster, Miguel de Val-Borro, Seth W. Digel, Peter Doherty, Richard Dubois, Gregory P. Dubois-Felsmann, Josef Durech, Frossie Economou, Michael Eracleous, Henry Ferguson, Enrique Figueroa, Merlin Fisher-Levine, Warren Focke, Michael D. Foss, James Frank, Michael D. Freemon, Emmanuel Gangler, Eric Gawiser, John C. Geary, Perry Gee, Marla Geha, Charles J. B. Gessner, Robert R. Gibson, D. Kirk Gilmore, Thomas Glanzman, William Glick, Tatiana Goldina, Daniel A. Goldstein,
25, 48
Iain Goodenow, Melissa L. Graham, William J. Gressler, Philippe Gris, Leanne P. Guy, Augustin Guyonnet, Gunther Haller, Ron Harris, Patrick A. Hascall, Justine Haupt, Fabio Hernandez, Sven Herrmann, Edward Hileman, Joshua Hoblitt, John A. Hodgson, Craig Hogan, Dajun Huang, Michael E. Huffer, Patrick Ingraham, Walter R. Innes, Suzanne H. Jacoby, Bhuvnesh Jain, Fabrice Jammes, James Jee, Tim Jenness, Garrett Jernigan, Darko Jevremovi´c, Kenneth Johns, Anthony S. Johnson, Margaret W. G. Johnson, R. Lynne Jones, Claire Juramy-Gilles, Mario Juri´c, Jason S. Kalirai, Nitya J. Kallivayalil, Bryce Kalmbach, Jeffrey P. Kantor, Pierre Karst, Mansi M. Kasliwal, Heather Kelly, Richard Kessler, Veronica Kinnison, David Kirkby, Lloyd Knox, Ivan V. Kotov, Victor L. Krabbendam, K. Simon Krughoff, Petr Kub´anek, John Kuczewski, Shri Kulkarni, John Ku, Nadine R. Kurita, Craig S. Lage, Ron Lambert,
2, 60
Travis Lange, J. Brian Langton, Laurent Le Guillou,
16, 11
Deborah Levine, Ming Liang, Kian-Tat Lim, Chris J. Lintott, Kevin E. Long, Margaux Lopez, Paul J. Lotz, Robert H. Lupton, Nate B. Lust, Lauren A. MacArthur, Ashish Mahabal, Rachel Mandelbaum, Darren S. Marsh, Philip J. Marshall, Stuart Marshall, Morgan May, Robert McKercher, Michelle McQueen, Joshua Meyers, Myriam Migliore, Michelle Miller, David J. Mills, Connor Miraval, Joachim Moeyens, David G. Monet, Marc Moniez, Serge Monkewitz, Christopher Montgomery, Fritz Mueller, Gary P. Muller, Freddy Mu˜noz Arancibia, Douglas R. Neill, Scott P. Newbry, Jean-Yves Nief, Andrei Nomerotski, Martin Nordby, Paul O’Connor, John Oliver,
41, 64
Scot S. Olivier, Knut Olsen, William O’Mullane, Sandra Ortiz, Shawn Osier, Russell E. Owen, Reynald Pain, Paul E. Palecek, John K. Parejko, James B. Parsons, Nathan M. Pease, J. Matt Peterson, John R. Peterson, Donald L. Petravick, M. E. Libby Petrick, Cathy E. Petry, Francesco Pierfederici, Stephen Pietrowicz, Rob Pike, Philip A. Pinto, Raymond Plante, Stephen Plate, Paul A. Price, Michael Prouza, Veljko Radeka, Jayadev Rajagopal, Andrew P. Rasmussen, Nicolas Regnault, Kevin A. Reil, David J. Reiss, Michael A. Reuter, Stephen T. Ridgway, Vincent J. Riot, Steve Ritz, Sean Robinson, William Roby, Aaron Roodman, Wayne Rosing, Cecille Roucelle, Matthew R. Rumore, Stefano Russo, Abhijit Saha, Benoit Sassolas, Terry L. Schalk, Pim Schellart,
7, 69
Rafe H. Schindler, Samuel Schmidt, Donald P. Schneider, Michael D. Schneider, William Schoening, German Schumacher,
2, 60
Megan E. Schwamb, Jacques Sebag, Brian Selvy, Glenn H. Sembroski, Lynn G. Seppala, Andrew Serio, Eduardo Serrano, Richard A. Shaw, Ian Shipsey, Jonathan Sick, Nicole Silvestri, Colin T. Slater, J. Allyn Smith, R. Chris Smith, Shahram Sobhani, Christine Soldahl, Lisa Storrie-Lombardi, Edward Stover, Michael A. Strauss, Rachel A. Street, Christopher W. Stubbs,
41, 64
Ian S. Sullivan, Donald Sweeney, John D. Swinbank,
1, 7
Alexander Szalay, Peter Takacs, Stephen A. Tether, Jon J. Thaler, John Gregg Thayer, Sandrine Thomas, Vaikunth Thukral, Jeffrey Tice, David E. Trilling, Max Turri, Richard Van Berg,
13, 52
Daniel Vanden Berk, Kurt Vetter, Francoise Virieux, Tomislav Vucina, William Wahl, a r X i v : . [ a s t r o - ph ] M a y Ivezi´c, Kahn, Tyson, Abel, Acosta, Allsman, Alonso, AlSayyad, Anderson, et al.
Lucianne Walkowicz,
77, 78
Brian Walsh, Christopher W. Walter, Daniel L. Wang, Shin-Yawn Wang, Michael Warner, Oliver Wiecha, Beth Willman,
2, 8
Scott E. Winters, David Wittman, Sidney C. Wolff, W. Michael Wood-Vasey, Xiuqin Wu, Bo Xin, Peter Yoachim, and Hu Zhan University of Washington, Dept. of Astronomy, Box 351580, Seattle, WA 98195 LSST Project Office, 950 N. Cherry Avenue, Tucson, AZ 85719 Kavli Institute for Particle Astrophysics and Cosmology, SLAC National Accelerator Laboratory, Stanford University, Stanford, CA94025 Physics Department, University of California, One Shields Avenue, Davis, CA 95616 Olympic College, 1600 Chester Ave., Bremerton, WA 98337-1699 Department of Physics, University of Oxford, Denys Wilkinson Building, Keble Road, Oxford, OX1 3RH, UK Department of Astrophysical Sciences, Princeton University, Princeton, NJ 08544 Steward Observatory, The University of Arizona, 933 N Cherry Ave., Tucson, AZ 85721 Giant Magellan Telescope Organization (GMTO), 465 N. Halstead Street, Suite 250, Pasadena, CA 91107 Laboratoire de l’Acc´el´erateur Lin´eaire, CNRS/IN2P3, Universit´e de Paris-Sud, B.P. 34, 91898 Orsay Cedex, France Laboratoire de Physique Nucl´eaire et des Hautes Energies, Universit´e Pierre et Marie Curie, Universit´e Paris Diderot, CNRS/IN2P3, 4place Jussieu, 75005 Paris, France AstroParticule et Cosmologie, Universit´e Paris Diderot, CNRS/IN2P3, CEA/lrfu, Observatoire de Paris, Sorbonne Paris Cit´e, 10, rueAlice Domon et L´eonie Duquet, Paris Cedex 13, France SLAC National Accelerator Laboratory, 2575 Sand Hill Rd, Menlo Park CA 94025 Laboratoire de Physique Subatomique et de Cosmologie, Universit´e Grenoble-Alpes, CNRS/IN2P3, 53 av. des Martyrs, 38026 Grenoblecedex, France Lawrence Livermore National Laboratory, 7000 East Avenue, Livermore, CA 94550 Sorbonne Universit´es, UPMC Univ Paris 06, UMR 7585, LPNHE, F-75005, Paris, France NCSA, University of Illinois at Urbana-Champaign, 1205 W. Clark St., Urbana, IL 61801 Brookhaven National Laboratory, Upton, NY 11973 Center for Urban Science & Progress, New York University, Brooklyn, NY 11021 Center for Cosmology & Particle Physics, New York University, New York, 10012 Oskar Klein Centre, Department of Physics, Stockholm University, SE 106 91 Stockholm, Sweden Universit´e Paris Diderot, Sorbonne Paris Cit´e, F-75013 Paris, France Center for Cosmology and Astro-Particle Physics, The Ohio State University, Columbus, OH 43210, USA Institute of Physics, Laboratory of Astrophysics, ´Ecole Polytechnique Fed`erale de Lausanne (EPFL), Observatoire de Sauverny, 1290Versoix, Switzerland Astronomy Department, University of California, 601 Campbell Hall, Berkeley, CA 94720 School of Physics, Astronomy and Computational Sciences, George Mason University, 4400 University Drive, Fairfax, VA 22030 Universit´e Grenoble-Alpes, Universit´e Savoie Mont Blanc, CNRS/IN2P3 Laboratoire d’Annecy-le-Vieux de Physique des Particules, 9Chemin de Bellevue – BP 110, 74940 Annecy-le-Vieux Cedex, France Department of Astronomy and Astrophysics, The Pennsylvania State University, 525 Davey Lab, University Park, PA 16802 Division of Geological and Planetary Sciences, California Institute of Technology, Pasadena, CA 91125 Center for Cosmology, University of California, Irvine, CA 92697 Laboratoire des Materiaux Avances (LMA), CNRS/IN2P3, Universit´e de Lyon, F-69622 Villeurbanne, Lyon, France Department of Computer Science, The University of Arizona, 1040 E 4th St, Tucson, AZ 85719 Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA 91109 Department of Physics, University of Arizona, 1118 E. Fourth Street, Tucson, AZ 85721 IPAC, California Institute of Technology, MS 100-22, Pasadena, CA 91125 Laboratoire Univers et Particules de Montpellier (LUPM), Universit´e de Montpellier, CNRS/IN2P3, Place Eug`ene Bataillon, 34095Montpellier, France Cook Astronomical Consulting, 220 Duxbury CT, San Ramon, CA 94583, USA Western Washington University, 516 High Street, Bellingham, WA 98225 Department of Physics and Astronomy, Purdue University, 525 Northwestern Ave., West Lafayette, IN 47907 University of Arizona, Tucson, AZ 85721 Department of Physics, Harvard University, 17 Oxford St, Cambridge MA 02138 Astronomical Institute, Charles University, Praha, Czech Republic Space Telescope Science Institute, 3700 San Martin Drive, Baltimore, MD 21218 Universit´e Clermont Auvergne, CNRS, Laboratoire de Physique de Clermont, F-63000 Clermont-Ferrand, France Department of Physics and Astronomy, Rutgers University, 136 Frelinghuysen Rd, Piscataway, NJ 08854 Smithsonian Astrophysical Observatory, 60 Garden St., Cambridge MA 02138 Astronomy Department, Yale University, New Haven, CT 06520 he Large Synoptic Survey Telescope Lawrence Berkeley National Laboratory, 1 Cyclotron Road, Berkeley, CA 94720, USA National Optical Astronomy Observatory, 950 N. Cherry Ave, Tucson, AZ 85719 CNRS, CC-IN2P3, 21 avenue Pierre de Coubertin, CS70202, 69627 Villeurbanne cedex, France Department of Astronomy and Astrophysics, University of Chicago, 5640 South Ellis Avenue, Chicago, IL 60637 Department of Physics & Astronomy, University of Pennsylvania, 209 South 33rd Street, Philadelphia, PA 19104-6396 Space Sciences Lab, University of California, 7 Gauss Way, Berkeley, CA 94720-7450 Astronomical Observatory, Volgina 7, P.O. Box 74, 11060 Belgrade, Serbia Department of Astronomy, University of Virginia, Charlottesville, VA 22904 Aix Marseille Univ, CNRS/IN2P3, CPPM, Marseille, France Astronomy Department, California Institute of Technology, 1200 East California Blvd., Pasadena CA 91125 Department of Physics and Astronomy, University of California, 4129 Frederick Reines Hall, Irvine, CA 92697 Institute of Physics, Academy of Sciences of the Czech Republic, Na Slovance 2, 182 21 Praha 8, Czech Republic Cerro Tololo InterAmerican Observatory, La Serena, Chile Longhorn Industries, Ellensburg, WA 98926 McWilliams Center for Cosmology, Department of Physics, Carnegie Mellon University, Pittsburgh, PA 15213, USA U.S. Naval Observatory Flagstaff Station, 10391 Naval Observatory Road, Flagstaff, AZ 86001 Department of Astronomy, Center for Astrophysics, Harvard University, 60 Garden St., Cambridge, MA 02138 Instituto de Radioastronom´ıa Milim´etrica, Av. Divina Pastora 7, N´ucleo Central, E-18012 Granada, Spain Google Inc., 1600 Amphitheatre Parkway Mountain View, CA 94043 Santa Cruz Institute for Particle Physics and Physics Department, University of California–Santa Cruz, 1156 High St., Santa Cruz,CA 95064 Las Cumbres Observatory, 6740 Cortona Dr. Suite 102, Santa Barbara, CA 93117 Department of Astrophysics/IMAPP, Radboud University Nijmegen, P.O. Box 9010, 6500 GL Nijmegen, The Netherlands Gemini Observatory, Northern Operations Center, 670 North A’ohoku Place, Hilo, HI 96720, USA Austin Peay State University, Clarksville, TN 37044 Belldex IT Consulting, Tucson, AZ 85742 Department of Physics and Astronomy, The John Hopkins University, 3701 San Martin Drive, Baltimore, MD 21218 University of Illinois, Physics and Astronomy Departments, 1110 W. Green St., Urbana, IL 61801 Department of Physics and Astronomy, Northern Arizona University, PO Box 6010, Flagstaff, AZ 86011, USA Saint Vincent College, Department of Physics, 300 Fraser Purchase Road, Latrobe, PA 15650 Library of Congress, 101 Independence Ave SE, Washington, DC 20540 The Adler Planetarium, 1300 South Lakeshore Ave, Chicago, IL 60605, USA Department of Physics, Duke University, Durham, NC 27708 Department of Physics and Astronomy, University of Pittsburgh, 3941 O’Hara Street, Pittsburgh PA 15260 Key Laboratory of Optical Astronomy, National Astronomical Observatories, Chinese Academy of Sciences, 20A Datun Road,Chaoyang District, Beijing 100012, China
ABSTRACTMajor advances in our understanding of the Universe frequently arise from dramatic improvementsin our ability to accurately measure astronomical quantities. Aided by rapid progress in informationtechnology, current sky surveys are changing the way we view and study the Universe. Next-generationsurveys will maintain this revolutionary progress. We describe here the most ambitious survey currentlyplanned in the optical, the Large Synoptic Survey Telescope (LSST). A vast array of science will beenabled by a single wide-deep-fast sky survey, and LSST will have unique survey capability in thefaint time domain. The LSST design is driven by four main science themes: probing dark energy anddark matter, taking an inventory of the Solar System, exploring the transient optical sky, and mappingthe Milky Way. LSST will be a large, wide-field ground-based system designed to obtain repeatedimages covering the sky visible from Cerro Pach´on in northern Chile. The telescope will have an 8.4m (6.5 m effective) primary mirror, a 9.6 deg field of view, and a 3.2 Gigapixel camera. The standardobserving sequence will consist of pairs of 15-second exposures in a given field, with two such visits ineach pointing in a given night to identify and constrain the orbits of asteroids. With these repeats,the LSST system is capable of imaging about 10,000 square degrees of sky in a single filter in threeclear nights. The typical 5 σ point-source depth in a single visit in r will be ∼ . Ivezi´c, Kahn, Tyson, Abel, Acosta, Allsman, Alonso, AlSayyad, Anderson, et al. will be contained within 30,000 deg with δ < +34 . ◦ , 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 devotedto a deep-wide-fast survey mode which will uniformly observe a 18,000 deg region about 800 times(summed over all six bands) during the anticipated 10 years of operations, and will yield a coaddedmap to r ∼ .
5. These data will result in databases including 20 billion galaxies and a similar numberof stars, and will serve the majority of the primary science programs. The remaining 10% of theobserving time will be allocated to special projects such as a Very Deep and Fast time domain survey,whose details are currently under discussion. We illustrate how the LSST science drivers led to thesechoices of system parameters, and describe the expected data products and their characteristics. Thegoal is to make LSST data products including a relational database of about 32 trillion observationsof 40 billion objects available to the public and scientists around the world – everyone will be able toview and study a high-definition color movie of the deep Universe.
Keywords: astronomical data bases: atlases, catalogs, surveys — Solar System — stars — the Galaxy— galaxies — cosmology INTRODUCTIONMajor advances in our understanding of the Universehave historically arisen from dramatic improvements inour ability to “see”. We have developed progressivelylarger telescopes over the past century, allowing us topeer further into space, and further back in time. Withthe development of advanced instrumentation – imagers,spectrographs, and polarimeters – we have been ableto parse radiation detected from distant sources overthe full electromagnetic spectrum in increasingly sub-tle ways. These data have provided the detailed infor-mation needed to construct physical models of planets,stars, galaxies, quasars, and larger structures, and toprobe the new physics of dark matter and dark energy.Until recently, most astronomical investigations havefocused on small samples of cosmic sources or individualobjects. This is because our largest telescope facilitiestypically had rather small fields of view, and those withlarge fields of view could not detect very faint sources.With all of our existing telescope facilities, we have stillsurveyed only a small fraction of the observable Universe(except when considering the most luminous quasars).Over the past two decades, however, advances in tech-nology have made it possible to move beyond the tradi-tional observational paradigm and to undertake large-scale sky surveys. As vividly demonstrated by sur-veys such as the Sloan Digital Sky Survey (SDSS; Yorket al. 2000), the Two Micron All Sky Survey (2MASS;Skrutskie et al. 2006), the Galaxy Evolution Explorer(GALEX; Martin et al. 2005), and Gaia (Gaia Collab-oration et al. 2016) to name but a few, sensitive andaccurate multi-color surveys over a large fraction of thesky enable an extremely broad range of new scientificinvestigations. These projects, based on a synergy ofadvances in telescope construction, detectors, and aboveall, information technology, have dramatically impacted nearly all fields of astronomy – and several areas of fun-damental physics. In addition, the world-wide atten-tion received by Sky in Google Earth (Scranton et al.2007), the World Wide Telescope , and the hundredsof thousands of volunteers classifying galaxies in theGalaxy Zoo project (Lintott et al. 2011) and its ex-tensions demonstrate that the impact of sky surveysextends far beyond fundamental science progress andreaches all of society.Motivated by the evident scientific progress enabledby large sky surveys, three nationally-endorsed reportsby the U.S. National Academy of Sciences (NationalResearch Council 2001, 2003a,b) concluded that a ded-icated ground-based wide-field imaging telescope withan effective aperture of 6–8 meters “is a high priorityfor planetary science, astronomy, and physics over thenext decade.” The Large Synoptic Survey Telescope(LSST) described here is such a system. Located onCerro Pach´on in northern Chile, the LSST will be alarge, wide-field ground-based telescope designed to ob-tain multi-band images over a substantial fraction of thesky every few nights. The survey will yield contiguousoverlapping imaging of over half the sky in six opticalbands, with each sky location visited close to 1000 timesover 10 years. The 2010 report “New Worlds, New Hori-zons in Astronomy and Astrophysics” by the NRC Com-mittee for a Decadal Survey of Astronomy and Astro-physics (National Research Council 2010) ranked LSSTas its top priority for large ground-based projects, andin May 2014 the National Science Board approved theproject for construction. As of this writing, the LSSTconstruction phase is close to the peak of activity. Af- http://worldwidetelescope.org/home he Large Synoptic Survey Telescope § § § §
5, and broad educational and societal im-pacts of the project in §
6. Concluding remarks are pre-sented in §
7. This publication will be maintained at thearXiv.org site FROM SCIENCE DRIVERS TO REFERENCEDESIGNThe most important characteristic that determinesthe speed at which a system can survey a given sky areato a given flux limit (i.e., its depth) is its ´etendue (orgrasp), the product of its primary mirror area and theangular area of its field of view (for a given set of ob-serving conditions, such as seeing and sky brightness).The effective ´etendue for LSST will be greater than 300m deg , which is more than an order of magnitudelarger than that of any existing facility. For example,the SDSS, with its 2.5-m telescope (Gunn et al. 2006)and a camera with 30 imaging CCDs (Gunn et al. 1998),has an effective ´etendue of only 5.9 m deg .The range of scientific investigations which will be en-abled by such a dramatic improvement in survey capa-bility is extremely broad. Guided by the community-wide input assembled in the report of the Science Work-ing Group of the LSST in 2004 (Science Working Groupof the LSST & Strauss 2004), the LSST is designed toachieve goals set by four main science themes:1. Probing Dark Energy and Dark Matter;2. Taking an Inventory of the Solar System;3. Exploring the Transient Optical Sky;4. Mapping the Milky Way.Each of these four themes itself encompasses a vari-ety of analyses, with varying sensitivity to instrumentaland system parameters. These themes fully exercise thetechnical capabilities of the system, such as photomet-ric and astrometric accuracy and image quality. About https://arxiv.org/abs/0805.2366
90% of the observing time will be devoted to a deep-wide-fast (main) survey mode. The working paradigmis that all scientific investigations will utilize a commondatabase constructed from an optimized observing pro-gram (the main survey mode), such as that discussed in § The Main Science Drivers
The main science drivers are used to optimize varioussystem parameters. Ultimately, in this high-dimensionalparameter space, there is a manifold defined by the to-tal project cost. The science drivers must both justifythis cost, as well as provide guidance on how to opti-mize various parameters while staying within the costenvelope.Here we summarize the dozen or so most importantinterlocking constraints on data and system propertiesplaced by the four main science themes:1. The depth of a single visit to a given field;2. Image quality;3. Photometric accuracy;4. Astrometric accuracy;5. Optimal exposure time;6. The filter complement;7. The distribution of revisit times (i.e., the cadenceof observations), including the survey lifetime;8. The total number of visits to a given area of sky;9. The coadded survey depth;10. The distribution of visits on the sky, and the totalsky coverage;11. The distribution of visits per filter; and12. Parameters characterizing data processing anddata access (such as the maximum time allowedafter each exposure to report transient sources,and the maximum allowed software contributionto measurement errors).
Ivezi´c, Kahn, Tyson, Abel, Acosta, Allsman, Alonso, AlSayyad, Anderson, et al.
We present a detailed discussion of how these science-driven data properties are transformed to system pa-rameters below.2.1.1.
Probing Dark Energy and Dark Matter
Current models of cosmology require the existence ofboth dark matter and dark energy to match observa-tional constraints (Riess et al. 2007; Komatsu et al. 2009;Percival et al. 2010; LSST Dark Energy Science Col-laboration 2012; Weinberg et al. 2015), and referencestherein). Dark energy affects the cosmic history of boththe Hubble expansion and mass clustering. Distinguish-ing competing models for the physical nature of dark en-ergy, or alternative explanations involving modificationsof the General Theory of Relativity, will require percentlevel measurements of both the cosmic expansion andthe growth of dark matter structure as a function ofredshift. Any given cosmological probe is sensitive to,and thus constrains degenerate combinations of, severalcosmological and astrophysical/systematic parameters.Therefore the most robust cosmological constraints arethe result of using interlocking combinations of probes.The most powerful probes include weak gravitationallens cosmic shear (WL), galaxy clustering and baryonacoustic oscillations (LSS), the mass function and clus-tering of clusters of galaxies, time delays in lensed quasarand supernova systems (SL), and photometry of typeIa supernovae (SN) – all as functions of redshift. Us-ing the cosmic microwave background fluctuations asthe normalization, the combination of these probes canyield the needed precision to distinguish among mod-els of dark energy (see e.g., Zhan 2006, and referencestherein). The challenge is to turn this available precisioninto accuracy, by careful modeling and marginalizationover a variety of systematic effects (see e.g., Krause &Eifler 2017).Meanwhile, there are a number of astrophysical probesof the fundamental properties of dark matter worth ex-ploring, including, for example, weak and strong lensingobservations of the mass distribution in galaxies and iso-lated and merging clusters, in conjunction with dynami-cal and X-ray observations (see e.g., Dawson et al. 2012;Newman et al. 2013; Rocha et al. 2013), the numbersand gamma-ray emission from dwarf satellite galaxies(see e.g., Hargis et al. 2014; Drlica-Wagner et al. 2015),the subtle perturbations of stellar streams in the MilkyWay halo by dark matter substructure (Belokurov &Koposov 2016), and massive compact halo object mi-crolensing (Alcock et al. 2001).Three of the primary Dark Energy probes, WL, LSSand SN, provide unique and independent constraints onthe LSST system design (SciBook Ch. 11–15). Weak lensing (WL) techniques can be used to map thedistribution of mass as a function of redshift and therebytrace the history of both the expansion of the Uni-verse and the growth of structure (e.g., Hu & Tegmark1999; for recent reviews see Kilbinger 2015; Mandel-baum 2017). Measurements of cosmic shear as a func-tion of redshift allow determination of angular distancesversus cosmic time, providing multiple independent con-straints on the nature of dark energy. These investiga-tions require deep wide-area multi-color imaging withstringent requirements on shear systematics in at leasttwo bands, and excellent photometry in at least fivebands to measure photometric redshifts (a requirementshared with LSS, and indeed all extragalactic sciencedrivers). The strongest constraints on the LSST im-age quality arise from this science program. In orderto control systematic errors in shear measurement, thedesired depth must be achieved with many short expo-sures (allowing for systematics in the measurement ofgalaxy shapes related to the PSF and telescope point-ing to be diagnosed and removed). Detailed simula-tions of weak lensing techniques show that imaging over ∼ ,
000 deg to a 5 σ point-source depth of r AB ∼ . he Large Synoptic Survey Telescope r ∼
24. Good image quality is required to separateSN photometrically from their host galaxies. Observa-tions in at least five photometric bands will allow properK-corrected light curves to be measured over a rangeof redshift. Carrying out these K-corrections requiresthat the calibration of the relative offsets in photomet-ric zero points between filters and the system responsefunctions, especially near the edges of bandpasses, beaccurate to about 1% (Wood-Vasey et al. 2007), sim-ilar to the requirements from photometric redshifts ofgalaxies. Deeper data ( r >
26) for small areas of thesky can extend the discovery of SN to a mean redshiftof 0.7 (from ∼ . z ∼ z (cid:38)
1) would improve con-straints on the properties of dark energy as a functionof redshift.Finally, there will be powerful cross checks and com-plementarities with other planned or proposed surveys,such as Euclid (Laureijs et al. 2011) and WFIRST(Spergel et al. 2015), which will provide wide-fieldoptical-IR imaging from space; DESI (Levi et al. 2013)and PFS (Takada et al. 2014), which will measure spec-troscopic BAO with millions of galaxies; and SKA (radio). Large survey volumes are key to probing dy-namical dark energy models (with sub-horizon darkenergy clustering or anisotropic stresses). The cross-correlation of the three-dimensional mass distribution –as probed by neutral hydrogen in CHIME (Newburghet al. 2014), HIRAX (Newburgh et al. 2016) or SKA,or galaxies in DESI and PFS – with the gravitationalgrowth probed by tomographic shear in LSST will be acomplementary way to constrain dark energy propertiesbeyond simply characterizing its equation of state andto test the underlying theory of gravity. Current and fu-ture ground-based CMB experiments, such as AdvancedACT (De Bernardis et al. 2016), SPT-3G (Benson et al.2014), Simons Observatory, and CMB Stage-4 (Abaza-jian et al. 2016), will also offer invaluable opportunitiesfor cross-correlations with secondary CMB anisotropies.2.1.2. Taking an Inventory of the Solar System The small-body populations in the Solar System,such as asteroids, trans-Neptunian objects (TNOs) andcomets, are remnants of its early assembly. The historyof accretion, collisional grinding, and perturbation byexisting and vanished giant planets is preserved in theorbital elements and size distributions of those objects.Cataloging the orbital parameters, size distributions,colors and light curves of these small-body populationsrequires a large number of observations in multiple fil-ters, and will lead to insights into planetary formationand evolution by providing the basis and constraintsfor new theoretical models. In addition, collisions in themain asteroid belt between Mars and Jupiter still occur,and occasionally eject objects on orbits that may placethem on a collision course with Earth. Studying theproperties of main belt asteroids at sub-kilometer sizesis important for linking the near-Earth Object (NEO)population with its source in the main belt. About 20%of NEOs, the potentially hazardous asteroids (PHAs),are in orbits that pass sufficiently close to Earth’s orbit,to within 0.05 AU, that perturbations on time scalesof a century can lead to the possibility of collision. InDecember 2005, the U.S. Congress directed NASA toimplement a survey that would catalog 90% of NEOswith diameters larger than 140 meters by 2020.Discovering and linking objects in the Solar Systemmoving with a wide range of apparent velocities (fromseveral degrees per day for NEOs to a few arc secondsper day for the most distant TNOs) places strong con-straints on the cadence of observations, requiring closelyspaced pairs of observations (two or preferably threetimes per lunation) in order to link detections unam-biguously and derive orbits (SciBook Ch. 5). Individualexposures should be shorter than about 30 seconds tominimize the effects of trailing for the majority of mov-ing objects. The images must be well sampled to en-able accurate astrometry, with absolute accuracy of atleast 0.1 arcsec in order to measure orbital parametersof TNOs with enough precision to constrain theoreti-cal models and enable prediction of occultations. Thephotometry should be better than 1–2% to measure as-teroids’ colors and thus determine their types. The dif-ferent filters should be observed over a short time spanto reduce apparent variations in color due to changes inobserving geometry, but should be repeated over manylunations in order to determine phase curves and allowshape modeling.The Congressional mandate can be fulfilled with a10-meter-class telescope equipped with a multi-gigapixel For details see http://neo.jpl.nasa.gov/neo/report2007.html
Ivezi´c, Kahn, Tyson, Abel, Acosta, Allsman, Alonso, AlSayyad, Anderson, et al. camera, and a sophisticated and robust data processingsystem (Ivezi´c et al. 2007a). The images should reach adepth of at least 24.5 (5 σ for point sources) in the r bandto reach high completeness down to the 140 m mandatefor NEOs. Such an instrument would probe the ∼ Exploring the Transient Optical Sky
Recent surveys have shown the power of measuringvariability of celestial sources for studying gravitationallensing, searching for supernovae, determining the phys-ical properties of gamma-ray burst sources, discoveringgravitational wave counterparts, probing the structureof active galactic nuclei, studying variable star popula-tions, discovering exoplanets, and many other subjectsat the forefront of astrophysics (SciBook Ch. 8; Lawet al. 2009; Djorgovski et al. 2012; Rowe et al. 2014).Time-domain science has diverse requirements fortransient and variable phenomena that are physicallyand phenomenologically heterogeneous. It requires largearea coverage to enhance the probability of detectingrare events; good image quality to enable differencing ofimages, especially in crowded fields; good time sampling,necessary to distinguish different types of variables andto infer their properties (e.g., determining the intrinsicpeak luminosity of Type Ia supernovae requires measur-ing their light curve shape); accurate color informationto classify variable objects; long term persistent obser-vations to characterize slow-evolving transients (e.g.,tidal disruption events, super luminous supernovae athigh redshift, and luminous blue variables); and rapiddata reduction, classification, and reporting to the com-munity to allow immediate follow-up with spectroscopy,further optical photometry, and imaging in other wave-bands.Wide area, dense temporal coverage to deep limitingmagnitudes will enable the discovery and analysis of rareand exotic objects such as neutron stars and black holebinaries, novae and stellar flares, gamma-ray bursts andX-ray flashes, active galactic nuclei, stellar disruptionsby black holes (Bloom et al. 2011; Gezari et al. 2012),and possibly new classes of transients, such as binarymergers of supermassive black holes (Shields & Bonning2008), chaotic eruptions on stellar surfaces (Arnett &Meakin 2011), and, further yet, completely unexpectedphenomena.Such a survey would likely detect microlensing by starsand compact objects in the Milky Way, but also in theLocal Group and perhaps beyond (de Jong et al. 2008).Given the duration of the LSST it will also be possible to detect the parallax microlensing signal of intermedi-ate mass black holes and measure their masses (Gould1992). It would open the possibility of discovering popu-lations of binaries and planets via transits (e.g., Beaulieuet al. 2006; Drake et al. 2010; Choi et al. 2013; Batistaet al. 2014), as well as obtaining spectra of lensed starsin distant galaxies.A deep and persistent survey will discover precursorsof explosive and eruptive transients, generate large sam-ples of transients whose study has thus far been limitedby small sample size (e.g., different subtypes of core col-lapse SN, Bianco et al. 2014.)Time series ranging between one minute and ten yearscadence should be probed over a significant fractionof the sky. The survey’s cadence will be sufficient,combined with the large coverage, to serendipitouslycatch very short-lived events, such as eclipses in ultra-compact double degenerate binary systems (Andersonet al. 2005), to constrain the properties of fast faint tran-sients (such as optical flashes associated with gamma-ray bursts; Bloom et al. 2008), to detect electromag-netic counterparts to gravitational wave sources (Nis-sanke et al. 2013; Scolnic et al. 2018) and to further con-strain the properties of new classes of transients discov-ered by programs such as the Deep Lens Survey (Beckeret al. 2004), the Catalina Real-time Transient Survey(Drake et al. 2009), the Palomar Transient Factory (Lawet al. 2009), and the Zwicky Transient Factory (Bellm2014). Observations over a decade will enable the studyof long period variables, intermediate mass black holes,and quasars (Kaspi et al. 2007; MacLeod et al. 2010;Graham et al. 2014; Chapline & Frampton 2016).The next frontier in this field will require measuringthe colors of fast transients, and probing variability atfaint magnitudes. Classification of transients in close-to-real time will require access to the full photometrichistory of the objects, both before and after the transientevent (e.g., Mahabal et al. 2011).2.1.4.
Mapping the Milky Way
A major challenge in extragalactic cosmology todayconcerns the formation of structure on sub-galacticscales, where baryon physics becomes important, andthe nature of dark matter may manifest itself in ob-servable ways (e.g. Weinberg et al. 2015). The MilkyWay and its environment provide a unique dataset forunderstanding the detailed processes that shape galaxyformation and for testing the small-scale predictions ofour standard cosmological model. New insights intothe nature and evolution of the Milky Way will re-quire wide-field surveys to constrain its structure andaccretion history. Further insights into the stellar pop- he Large Synoptic Survey Telescope ∼
100 kpc(Ivezi´c et al. 2004) with main-sequence stars, the totalcoadded depth must reach r >
27, with a similar depthin the g band. The metallicity distribution of stars canbe studied photometrically in the Sgr tidal stream (e.g.,see Majewski et al. 2003; Chou et al. 2007) and otherhalo substructures ( ∼
30 kpc, Carollo et al. 2007), yield-ing new insights into how they formed. Our ability tomeasure these metallicities is limited by the coaddeddepth in the u band; to probe the outer parts of thestellar halo, one must reach u ∼ .
5. To detect RRLyrae stars beyond the Galaxy’s tidal radius at ∼ r ∼ . − (comparable with the accu-racy of large-scale radial velocity surveys), the propermotion accuracy should be 0.2 mas yr − or better. Thisis the same accuracy as will be delivered by the Gaia mis-sion (Perryman et al. 2001; de Bruijne 2012) at its faintlimit ( r ∼ σ ) are required over 10 years. To achievethe required proper motion and parallax accuracy withan assumed astrometric accuracy of 10 mas per obser-vation per coordinate, approximately 1,000 separate ob-servations are required. This requirement for a largenumber of observations is similar to that from minimiz-ing systematics in weak lensing observations ( § A Summary and Synthesis of Science-drivenConstraints on Data Properties
The goals of all the science programs discussed above(and many more, of course) can be accomplished by sat-isfying the minimal constraints listed below. For a moreelaborate listing of various constraints, including de- http://sci.esa.int/gaia/ Figure 1.
The image quality distribution measured at theCerro Pach´on site using a differential image motion monitor(DIMM) at λ = 500 nm, and corrected using an outer scaleparameter of 30 m over an 8.4 m aperture. For details aboutthe outer scale correction see Tokovinin (2002). The observeddistribution is well described by a log-normal distribution,with the parameters shown in the figure. tailed specification of various probability density distri-bution functions, please see the LSST Science Require-ments Document (Ivezi´c & The LSST Science Collabo-ration 2011) and the LSST Science Book (LSST ScienceCollaboration et al. 2009).1. The single visit depth should reach r ∼ .
5. Thislimit is primarily driven by the search for NEOs,variable sources (e.g., SN, RR Lyrae stars), andby proper motion and trigonometric parallax mea-surements for stars. Indirectly, it is also drivenby the requirements on the coadded survey depthand the minimum number of exposures requiredby WL science. We plan to split a single visit intotwo exposures of equal length to identify and re-move cosmic rays.2.
Image quality should maintain the limit set by theatmosphere (the median free-air seeing is 0.65 arc-sec in the r band at the chosen site, see Fig. 1), andnot be degraded appreciably by the hardware. Inaddition to stringent constraints from weak lens-ing, good image quality is driven by the requiredsurvey depth for point sources and by image dif-ferencing techniques.3. Photometric repeatability should achieve 5 mmagprecision at the bright end, with zeropoint stabil-0
Ivezi´c, Kahn, Tyson, Abel, Acosta, Allsman, Alonso, AlSayyad, Anderson, et al. ity across the sky of 10 mmag and band-to-bandcalibration errors not larger than 5 mmag. Theserequirements are driven by the need for high pho-tometric redshift accuracy, the separation of stellarpopulations, detection of low-amplitude variableobjects (such as eclipsing planetary systems), andthe search for systematic effects in type Ia super-nova light curves.4.
Astrometric precision should maintain the limitset by the atmosphere, of about 10 mas per visitat the bright end (on scales below 20 arcmin). Thisprecision is driven by the desire to achieve a propermotion accuracy of 0.2 mas yr − and parallax ac-curacy of 1.0 mas over the course of a 10-year sur-vey (see § The single visit exposure time should be less thanabout a minute to prevent trailing of fast movingobjects and to aid control of various systematiceffects induced by the atmosphere. It should belonger than ∼
20 seconds to avoid significant effi-ciency losses due to finite readout, slew time, andread noise. As described above, we are planningto split each visit into two exposures.6.
The filter complement should include at least sixfilters in the wavelength range limited by atmo-spheric absorption and silicon detection efficiency(320–1050 nm), with roughly rectangular filtersand no large gaps in the coverage, in order toenable robust and accurate photometric redshiftsand stellar typing. An SDSS-like u band (Fukugitaet al. 1996) is extremely important for separatinglow-redshift quasars from hot stars, and for es-timating the metallicities of F/G main sequencestars. A bandpass with an effective wavelengthof about 1 micron would enable studies of sub-stellar objects, high-redshift quasars (to redshiftsof ∼ The revisit time distribution should enable deter-mination of orbits of Solar System objects andsample SN light curves every few days, while ac-commodating constraints set by proper motionand trigonometric parallax measurements.8.
The total number of visits of any given area of sky,when accounting for all filters, should be of theorder of 1,000, as mandated by WL science, thesearch for NEOs, and proper motion and trigono-metric parallax measurements. Studies of tran-sient sources also benefit from a large number ofvisits. 9.
The coadded survey depth should reach r ∼ . The distribution of visits per filter should en-able accurate photometric redshifts, separation ofstellar populations, and sufficient depth to en-able detection of faint extremely red sources (e.g.,brown dwarfs and high-redshift quasars). Detailedsimulations of photometric redshift uncertaintiessuggest roughly similar number of visits amongbandpasses (but because the system throughputand atmospheric properties are wavelength depen-dent, the achieved depths are different in differentbands). The adopted time allocation (see Table 1)includes a slight preference to the r and i bandsbecause of their dominant role in star/galaxy sep-aration and weak lensing measurements.11. The distribution of visits on the sky should ex-tend over at least ∼ to obtain the re-quired number of galaxies for WL studies, withattention paid to include “special” regions such asthe Ecliptic and Galactic planes, and the Largeand Small Magellanic Clouds (if in the SouthernHemisphere). For comparison, the full area thatcan be observed at airmass less than 2.0 from anymid-latitude site is about 30,000 deg .12. Data processing, data products and data access should result in data products that approach thestatistical uncertainties in the raw data; i.e., theprocessing must be close to optimal. To enablefast and efficient response to transient sources, theprocessing latency for variable sources should beless than a minute, with a robust and accuratepreliminary characterizationof all reported variables.Remarkably, even with these joint requirements, noneof the individual science programs is severely over-designed, i.e., despite their significant scientific diver-sity, these programs are highly compatible in terms ofdesired data characteristics. Indeed, any one of the fourmain science drivers could be removed, and the remain-ing three would still yield very similar requirements formost system parameters. As a result, the LSST systemcan adopt a highly efficient survey strategy in which a single dataset serves most science programs (insteadof science-specific surveys executed in series). One canview this project as massively parallel astrophysics . Thevast majority (about 90%) of the observing time will be he Large Synoptic Survey Telescope Table 1.
The LSST Baseline Design and Survey ParametersQuantity Baseline Design SpecificationOptical Config. 3-mirror modified Paul-BakerMount Config. Alt-azimuthFinal f-ratio, aperture f/1.234, 8.4 mField of view, ´etendue 9.6 deg , 319 m deg Plate Scale 50.9 µ m/arcsec (0.2” pix)Pixel count 3.2 GigapixWavelength Coverage 320 – 1050 nm, ugrizy Single visit depths, design a b c
56, 80, 184, 184, 160, 160Final (coadded) depths d a Design specification from the Science Requirements Document(SRD; Ivezi´c & The LSST Science Collaboration 2011) for 5 σ depths for point sources in the ugrizy bands, respectively. Thelisted values are expressed on the AB magnitude scale, and cor-respond to point sources and fiducial zenith observations (about0.2 mag loss of depth is expected for realistic airmass distribu-tions, see Table 2 for more details). b Minimum specification from the Science Requirements Docu-ment for 5 σ depths. c An illustration of the distribution of the number of visits as afunction of bandpass, taken from Table 24 in the SRD. d Idealized depth of coadded images, based on design specificationfor 5 σ depth and the number of visits in the penultimate row(taken from Table 24 in the SRD). devoted to a deep-wide-fast survey mode of the sort wehave just described, with the remaining 10% allocatedto special programs which will also address multiple sci-ence goals. Before describing these surveys in detail, wediscuss the main system parameters.2.2. The Main System Design Parameters
Given the minimum science-driven constraints on thedata properties listed in the previous section, we nowdiscuss how they are translated into constraints on themain system design parameters: the aperture size, thesurvey lifetime, the optimal exposure time, and the filtercomplement. 2.2.1.
The Aperture Size
The product of the system’s ´etendue and the surveylifetime, for given observing conditions, determines thesky area that can be surveyed to a given depth. TheLSST field-of-view area is maximized to its practicallimit, ∼
10 deg , determined by the requirement that the delivered image quality be dominated by atmosphericseeing at the chosen site (Cerro Pach´on in NorthernChile). A larger field-of-view would lead to unaccept-able deterioration of the image quality. This constraintleaves the primary mirror diameter and survey lifetimeas free parameters. The adopted survey lifetime of 10years is a compromise between a shorter time that leadsto an excessively large and expensive mirror (15 m fora 3 year survey and 12 m for a 5 year survey) and notas effective proper motion measurements, and a smallertelescope that would require more time to complete thesurvey, with the associated increase in operations cost.The primary mirror size is a function of the requiredsurvey depth and the desired sky coverage. By andlarge, the anticipated science outcome scales with thenumber of detected sources. For practically all astro-nomical source populations, in order to maximize thenumber of detected sources, it is more advantageous tomaximize the area first, and then the detection depth .For this reason, the sky area for the main survey is max-imized to its practical limit, 18,000 deg , determined bythe requirement to avoid airmasses less than 1.5, whichwould substantially deteriorate the image quality andthe survey depth (see eq. 6).With the adopted field-of-view area, the sky cover-age and the survey lifetime fixed, the primary mirrordiameter is fully driven by the required survey depth.There are two depth requirements: the final (coadded)survey depth, r ∼ .
5, and the depth of a single visit, r ∼ .
5. The two requirements are compatible if thenumber of visits is several hundred per band, which isin good agreement with independent science-driven re-quirements on the latter.The required coadded survey depth provides a directconstraint, independent of the details of survey execu-tion such as the exposure time per visit, on the minimumeffective primary mirror diameter of 6.4 m, as illustratedin Fig. 2. 2.2.2.
The Optimal Exposure Time
The single visit depth depends on both the primarymirror diameter and the chosen exposure time, t vis . Inturn, the exposure time determines the time intervalto revisit a given sky position and the total number of If the total exposure time is doubled and used to double thesurvey area, the number of sources increases by a factor of two.If the survey area is kept fixed, the increased exposure time willresult in ∼ N ) = C + k ∗ m , the number of sourceswill increase by more than a factor of two only if k > .
75. Apartfrom z < k at most 0.6(the Euclidean value), and faint stars and galaxies have k < . Ivezi´c, Kahn, Tyson, Abel, Acosta, Allsman, Alonso, AlSayyad, Anderson, et al. visits, and each of these quantities has its own sciencedrivers. We summarize these simultaneous constraintsin terms of the single-visit exposure time: • The single-visit exposure time should not be longerthan about a minute to prevent trailing of fast So-lar System moving objects, and to enable efficientcontrol of atmospheric systematics. • The mean revisit time (assuming uniform cadence)for a given position on the sky, n , scales as n = (cid:18) t vis
10 sec (cid:19) (cid:18) A sky ,
000 deg (cid:19) (cid:18)
10 deg A FOV (cid:19) days , (1)where two visits per night are assumed (requiredfor efficient detection of Solar System objects, seebelow), and the losses for realistic observing condi-tions have been taken into account (with the aid ofthe Operations Simulator described below). Sci-ence drivers such as supernova light curves andmoving objects in the Solar System require that n < t vis <
40 seconds forthe nominal values of A sky and A F OV . • The number of visits to a given position on the sky, N visit , with losses for realistic observing conditionstaken into account, is given by N visit = (cid:18) n (cid:19) (cid:18) T
10 yr (cid:19) . (2)The requirement N visit >
800 again implies that n < t vis <
40 seconds if the survey lifetime, T is about 10 years. • These three requirements place a firm upper limiton the optimal visit exposure time of t vis < t vis due to fi-nite detector read-out and telescope slew time (thelongest acceptable read-out time is set to 2 sec-onds, the shutter open-and-close time is 2 seconds,and the slew and settle time is set to 5 seconds, in-cluding the read-out time for the second exposurein a visit): (cid:15) = (cid:18) t vis t vis + 9 sec (cid:19) . (3)To maintain efficiency losses below ∼
30% (i.e., atleast below the limit set by the weather patterns),and to minimize the read noise impact, t vis > Figure 2.
The coadded depth in the r band (AB magni-tudes) vs. the effective aperture and the survey lifetime. Itis assumed that 22% of the total observing time (correctedfor weather and other losses) is allocated for the r band, andthat the ratio of the surveyed sky area to the field-of-viewarea is 2,000. Taking these constraints simultaneously into account,as summarized in Fig. 3, yielded the following referencedesign:1. A primary mirror effective diameter of ∼ ∼ (cid:15) = 77%.3. A revisit time of 3 days on average for 10,000 deg of sky, with two visits per night.To summarize, the chosen primary mirror diameter isthe minimum diameter that simultaneously satisfies thedepth ( r ∼ . r ∼ . System Design Trade-offs he Large Synoptic Survey Telescope Figure 3.
The single-visit depth in the r band (5 σ detectionfor point sources, AB magnitudes) vs. revisit time, n (days),as a function of the effective aperture size. With a cover-age of 10,000 deg in two bands, the revisit time directlyconstrains the visit exposure time, t vis = 10 n seconds. Inaddition to direct constraints on optimal exposure time, t vis is also driven by requirements on the revisit time, n , the to-tal number of visits per sky position over the survey lifetime, N visit , and the survey efficiency, (cid:15) (see eqs.1-3). Note thatthese constraints result in a fairly narrow range of allowed t vis for the main deep-wide-fast survey. We note that the Pan-STARRS project (Kaiser et al.2002, 2010), with similar science goals as LSST, envi-sions a distributed aperture design, where the total sys-tem ´etendue is a sum of ´etendue values for an arrayof small 1.8 m telescopes . Similarly, the LSST systemcould perhaps be made as two smaller copies with 6mmirrors, or 4 copies with 4m mirrors, or 16 copies with2m mirrors. Each of these clones would have to haveits own 3 Gigapixel camera (see below), and given theadded risk and complexity (e.g., maintenance, data pro-cessing), the monolithic design seems advantageous fora system with such a large ´etendue as LSST.It is informative to consider the tradeoffs that wouldbe required for a system with a smaller aperture, if thescience requirements were to be maintained. For thiscomparison, we consider a four-telescope version of the The first of these telescopes, PS1, has been operational forsome time (Chambers et al. 2016), and has an ´etendue 1/24 th that of LSST. Pan-STARRS survey (PS4). With an ´etendue about 6times smaller than that of LSST (effective diameters of6.4 m and 3.0 m, and a field-of-view area of 9.6 deg vs. 7.2 deg ), and all observing conditions being equal,the PS4 system could in principle use a cadence identi-cal to that of LSST. The main difference in the datasetswould be a faint limit shallower by about 1 mag in agiven survey lifetime. As a result, for Euclidean popu-lations the sample sizes would go down by a factor of 4,while for populations of objects with a shallower slopeof the number-magnitude relation (e.g., galaxies aroundredshift of 1) the samples would be smaller by a fac-tor 2–3. The distance limits for nearby sources, suchas Milky Way stars, would drop to 60% of their corre-sponding LSST values, and the NEO completeness levelmandated by the U.S. Congress would not be reached.If instead the survey coadded depth were to be main-tained, then the survey sky area would have to be 6 timessmaller ( ∼ ). If the survey single-visit depthwere to be maintained, then the exposure time wouldhave to be about 6 times longer (ignoring the slight dif-ference in the field-of-view area and simply scaling bythe ´etendue ratio), resulting in non-negligible trailinglosses for Solar System objects, and either i) a factorof six smaller sky area observed within n = 3 days,or ii) the same sky area revisited every n = 18 days.Given these conflicts, one solution would be to split theobserving time and allocate it to individual specializedprograms (e.g., large sky area vs. deep coadded data vs.deep single-visit data vs. small n data, etc.), as is beingdone by the PS1 Consortium .In summary, given the science requirements as statedhere, there is a minimum ´etendue of ∼
300 deg m whichenables our seemingly disparate science goals to be ad-dressed with a single dataset. A system with a smaller´etendue would require separate specialized surveys toaddress the science goals, which results in a loss of sur-veying efficiency . The LSST is designed to reach thisminimum ´etendue for the science goals stated in its Sci-ence Requirements Document.2.4. The Filter Complement
The LSST filter complement ( ugrizy , see Fig. 4) ismodeled after the Sloan Digital Sky Survey (SDSS)system (Fukugita et al. 1996) because of its demon-strated success in a wide variety of applications, includ- The converse is also true: for every ´etendue there is a set ofoptimal science goals that such a system can address with a highefficiency. Ivezi´c, Kahn, Tyson, Abel, Acosta, Allsman, Alonso, AlSayyad, Anderson, et al.
300 400 500 600 700 800 900 1000 1100Wavelength (nm)0.00.20.40.60.81.0 T h r o u g h p u t ( - ) Airmass 1.2 u g r i z y
Figure 4.
The LSST bandpasses. The vertical axis showsthe total throughput. The computation includes the atmo-spheric transmission (assuming an airmass of 1.2, dottedline), optics, and the detector sensitivity. ing photometric redshifts of galaxies (Budav´ari et al.2003), separation of stellar populations (Lenz et al. 1998;Helmi et al. 2003), and photometric selection of quasars(Richards et al. 2002; Ross et al. 2012). The extension ofthe SDSS system to longer wavelengths (the y band at ∼ ugrizy system with a filter complementthat includes only five filters. For example, each filterwidth could be increased by 20% over the same wave-length range (neither a shorter wavelength range, norgaps in the wavelength coverage are desirable options),but this option is not satisfactory. Placing the red edgeof the u band blueward of the Balmer break allows op-timal separation of stars and quasars, and the telluricwater absorption feature at 9500 ˚A effectively defines theblue edge of the y band. Of the remaining four filters( griz ), the g band is already quite wide. As a last op-tion, the riz bands could be redesigned as two widerbands. However, this option is also undesirable becausethe r and i bands are the primary bands for weak lensingstudies and for star/galaxy separation, and chromaticatmospheric refraction would worsen the point spreadfunction for a wider bandpass.2.5. The Calibration Methods
Wavelength (Angstrom) F l u x ( e r g s / c m / s / A n g s t r o m ) x1e-14 Figure 5.
An example of determination of the atmosphericopacity by simultaneously fitting a three-parameter stellarmodel SED (Kurucz 1979) and six physical parameters ofa sophisticated atmospheric model (MODTRAN, Andersonet al. 1999) to an observed F-type stellar spectrum ( F λ ). Theblack line is the observed spectrum and the red line is thebest fit. Note that the atmospheric water feature around0.9–1.0 µ m is exquisitely well fit. The components of thebest-fit atmospheric opacity are shown in Fig. 6. Adaptedfrom Burke et al. (2010). Figure 6.
The components of the best-fit atmospheric opac-ity used to model the observed stellar spectrum shown inFig. 5. The atmosphere model (MODTRAN, Anderson et al.1999) includes six components: water vapor (blue), oxy-gen and other trace molecules (green), ozone (red), Rayleighscattering (cyan), a gray term with a transmission of 0.989(not shown) and an aerosol contribution proportional to λ − and extinction of 1.3% at λ =0.675 µ m (not shown). Theblack line shows all six components combined. Adapted fromBurke et al. (2010). he Large Synoptic Survey Telescope R ∼ µ m) than shown in these figures,using a slitless spectrograph and an LSST corner-raftCCD. Celestial spectrophotometric standard stars canbe used as a separate means of photometric calibration,albeit only through the comparison of band-integratedfluxes with synthetic photometry calculations.A similar calibration process has been undertaken bythe Dark Energy Survey (DES) team, which has beenapproaching a calibration precision of 5 mmag (Burkeet al. 2018).SDSS, PS1, and DES data taken in good photomet-ric conditions have approached the LSST requirement of1% photometric calibration (Padmanabhan et al. 2008;Schlafly et al. 2012; Burke et al. 2018), although mea-surements with ground-based telescopes typically pro-duce data with errors a factor of two or so larger. Anal-ysis of repeated SDSS scans obtained in varying ob-serving conditions demonstrates that data obtained innon-photometric conditions can also be calibrated withsufficient accuracy (Ivezi´c et al. 2007b), as long as high-quality photometric data also exist in the region. TheLSST calibration plan builds on this experience gainedfrom the SDSS and other surveys.The planned calibration process decouples the estab-lishment of a stable and uniform internal relative cali- bration from the task of assigning absolute optical fluxto celestial objects.Celestial sources will be used to refine the internalphotometric system and to monitor stability and uni-formity of the photometric data. We expect to use GaiaCollaboration et al. (2016) photometry, utilising the BP and RP photometric measurements as well as the Gmagnitudes; for a subset of stars ( e.g. F-subdwarfs)we expect to be able to transfer this rigid photomet-ric system above the atmosphere to objects observedby LSST. There will be >
100 main-sequence stars with17 < r <
20 per detector (14 ×
14 arcmin ) even athigh Galactic latitudes. Standardization of photomet-ric scales will be achieved through direct observation ofstars with well-understood spectral energy distributions(SEDs), in conjunction with the in-dome calibration sys-tem and the atmospheric transmission spectra.Astrometric calibration will be based on the resultsfrom the Gaia mission (Gaia Collaboration et al. 2016),which will provide numerous high-accuracy astrometricstandards in every LSST field.2.6. The LSST Reference Design
We briefly describe the reference design for the mainLSST system components. Detailed discussion of theflow-down from science requirements to system designparameters, and extensive system engineering analysiscan be found in the LSST Science Book (Ch. 2–3).2.6.1.
Telescope and Site
The large LSST ´etendue is achieved in a novel three-mirror design (modified Paul-Baker Mersenne-Schmidtsystem; Angel et al. 2000) with a very fast f /1.234 beam.The optical design has been optimized to yield a largefield of view (9.6 deg ), with seeing-limited image qual-ity, across a wide wavelength band (320–1050 nm). In-cident light is collected by an annular primary mirror,having an outer diameter of 8.4 m and inner diameterof 5.0 m, creating an effective filled aperture of ∼ Ivezi´c, Kahn, Tyson, Abel, Acosta, Allsman, Alonso, AlSayyad, Anderson, et al.
Figure 7.
The LSST baseline optical design (modified three-mirror Paul-Baker) with its unique monolithic mirror: theprimary and tertiary mirrors are positioned such that theyform a continuous compound surface, allowing them to bepolished from a single substrate.
Figure 8.
The polishing of the primary-tertiary mirror pairat the Richard F. Caris Mirror Lab at the University of Ari-zona Tucson. diameter, thus making it possible to fabricate the mir-ror pair from a single monolithic blank using spin-castborosilicate technology. The secondary mirror is fab-ricated from a thin 100 mm thick meniscus substrate,made from Corning’s ultra-low expansion material. Allthree mirrors will be actively supported to control wave-front distortions introduced by gravity and environmen-tal stresses on the telescope. The primary-tertiary mir-
Figure 9.
The baseline design for the LSST telescope. Thesmall focal ratio allows for a very squat telescope, and thusa very stiff structure. ror was cast and polished by the Richard F. Caris MirrorLab at the University of Arizona in Tucson before beinginspected and accepted by LSST in April 2015 (Araujo-Hauck et al. 2016). The primary-tertiary mirror cell wasfabricated by CAID in Tucson and is undergoing accep-tance tests. The integration of the actuators and finaltests with the mirror is scheduled for early 2018.The LSST Observing Facility (Fig. 10), consisting ofthe telescope enclosure and summit support building, isbeing constructed atop Cerro Pach´on in northern Chile,sharing the ridge with the Gemini South and SOAR tele-scopes (latitude: S 30 ◦ (cid:48) (cid:48)(cid:48) ; longitude: W 70 ◦ (cid:48) (cid:48)(cid:48) ; elevation: 2652 m; Mamajek 2012). The tele-scope enclosure houses a compact, stiff telescope struc-ture (see Fig. 9) atop a 15 m high concrete pier with afundamental frequency of 8 Hz, that is crucial for achiev-ing the required fast slew-and-settle times. The height ofthe pier was set to place the telescope above the degrad-ing effects of the turbulent ground layer. Capping thetelescope enclosure is a 30 m diameter dome with exten-sive ventilation to reduce dome seeing and to maintaina uniform thermal environment over the course of thenight. Furthermore, the summit support building hasbeen oriented with respect to the prevailing winds toshed its turbulence away from the telescope enclosure.The summit support building includes a coating cham-ber for recoating the three LSST mirrors and clean roomfacilities for maintaining and servicing the camera.2.6.2. Camera Coordinates listed in older versions of this paper were incor-rect. We thank E. Mamajek for pointing out this error to us. he Large Synoptic Survey Telescope Figure 10.
Top: artist’s rendering of the dome enclosurewith the attached summit support building on Cerro Pach´on.The LSST auxiliary calibration telescope is shown on an ad-jacent rise to the right. Bottom: Photograph of the LSSTObservatory as of Summer 2017. Note the different perspec-tive from the artist’s rendering. The main LSST telescopebuilding is on the right, waiting for the dome to be installed.The auxiliary telescope building is on the left with its domebeing installed.
The LSST camera provides a 3.2 Gigapixel flat focalplane array, tiled by 189 4K ×
4K CCD science sensorswith 10 µ m pixels (see Figs. 11 and 12). This pixel countis a direct consequence of sampling the 9.6 deg field-of-view (0.64 m diameter) with 0.2 × pixels(Nyquist sampling in the best expected seeing of ∼ × Figure 11.
A cutaway view of LSST camera. Not shownare the shutter, which is positioned between the filter andlens L3, and the filter exchange system. in an optics structure at the front of the camera body,which also contains a mechanical shutter, and a carouselassembly that holds five large optical filters. The sixthoptical filter can replace any of the five via a procedureaccomplished during daylight hours.Each of the 21 rafts will host 3 front end electronicboards (REB) operating in the cryostat (at − ◦ C),that read in parallel a total of 9 ×
16 segments per CCD(144 video channels reading one million pixels each).This very high parallelization is the key to allow for afast readout (2 seconds) of the entire focal plane. Toreach this performance with a reasonably-sized board,a special low-noise ( < < Data Management
The rapid cadence and scale of the LSST observingprogram will produce approximately 15 TB per nightof raw imaging data (about 20 TB with calibrationexposures). As with all large modern surveys, the largedata volume, the real-time aspects, and the complexityof processing involved requires that the survey itselftake on the task of fully reducing the data. The datacollected by the LSST system will be automatically re-duced to scientifically useful catalogs and images by theLSST Data Management (DM; Juri´c et al. 2015) system. For comparison, the volume of all imaging data collected overa decade by the SDSS-I/II projects and published in SDSS DataRelease 7 (Abazajian et al. 2009) is approximately 16 TB. Ivezi´c, Kahn, Tyson, Abel, Acosta, Allsman, Alonso, AlSayyad, Anderson, et al.
Figure 12.
The LSST Camera focal plane array. Each cyansquare represents one 4K ×
4K pixel sensor. Nine sensors areassembled into a raft; the 21 rafts are outlined in red. Thereare 189 science sensors, for a total of 3.2 gigapixels. Alsoshown are the four corner rafts, where the guide sensors andwavefront sensors are located.
Figure 13.
The LSST Camera raft module, correspondingto the red squares in Fig. 12, with 9 sensors, integrated elec-tronics, and thermal connections. Raft modules are designedto be replaceable.
The detailed outputs of the LSST Data Managementsystem are described in § • Process, in real time, the incoming stream ofimages generated by the camera system duringobserving by archiving raw images, generatingalerts to new sources or sources whose propertieshave changed, and updating the relevant catalogs(Prompt products; § • Process each night’s data during the day and de-termine or refine orbits for all asteroids found inthe imaging. • Periodically process the accumulated survey datato provide a uniform photometric and astrometriccalibration, measure the properties of all detectedobjects, and characterize objects based on theirtime-dependent behavior. The results of such aprocessing run form a
Data Release (DR), whichis a static, self-consistent dataset suitable for usein performing scientific analyses of LSST data andpublication of the results (the data release prod-ucts; § • Facilitate the creation of data products gener-ated by the science community, by providing suit-able software, application programming interfaces(APIs), and computing infrastructure at the LSSTdata access centers. • Make all LSST data available through an inter-face that utilizes community-based standards tothe maximum possible extent. Provide enoughprocessing, storage, and network bandwidth to en-able user analyses of the data without the need forpetabyte-scale data transfers.Over the ten years of LSST operations and 11 datareleases, this processing will result in a cumulative pro-cessed data size approaching 500 petabytes (PB) forimaging, and over 50 PB for the catalog databases. Thefinal data release catalog database alone is expected tobe approximately 15 PB in size.The DM system will span four key facilities on threecontinents: the Summit Facility on Cerro Pach´on inChile (where the initial detector cross-talk correctionwill be performed); the Base Facility in La Serena, Chile(which will serve as a retransmission for data uploadsto North America, as well as the Data Access Centerfor the Chilean community); the Data Processing and he Large Synoptic Survey Telescope Figure 14.
The LSST data flow from the mountain facilities in Chile to the data access center and processing center in theU.S., and the satellite processing center in France.
Archiving Facility at the National Center for Supercom-puting Applications (NCSA) in Champaign-Urbana, IL;and the Satellite Processing Facility at CC-IN2P3 inLyon, France. All real-time data processing and halfthe data release product processing will take place atthe Data Processing and Archiving Facility, which willalso serve as the Data Access Center for the US commu-nity. The other half of the data release processing willbe done at CC-IN2P3, which will also have the role of“Long-term Storage” facility.The data will be transported between the centers overexisting and new high-speed optical fiber links fromSouth America to the U.S. (see Fig. 14). The data pro-cessing center demands stable, well-tested technology toensure smooth operations. Hence, while LSST is mak-ing a novel use of advances in information technology,it is not pushing the expected technology to the limit,reducing the overall risk to the project.2.6.4.
The LSST software stack
The
LSST Software Stack is the data processing andanalysis system developed by the LSST Project to en-able LSST survey data reduction and delivery. It com-prises all science pipelines needed to accomplish LSSTdata processing tasks (e.g., calibration, single frameprocessing, coaddition, image differencing, multi-epochmeasurement, asteroid orbit determination, etc.), the necessary data access and orchestration middleware, aswell as the database and user interface components.Algorithm development for the LSST software buildson the expertise and experience of prior large astronom-ical surveys (including SDSS, Pan-STARRS, DES, Su-perMACHO, ESSENCE, DLS, CFHTLS, and UKIDSS).The pipelines written for these surveys have demon-strated that it is possible to carry out largely au-tonomous data reduction of large datasets, automateddetection of sources and objects, and the extractionof scientifically useful characteristics of those objects.While firmly footed in this prior history, the LSST soft-ware stack has largely been written anew, for reasonsof performance, extendability, and maintainability. AllLSST codes have been designed and implemented follow-ing software engineering best practices, including mod-ularity, clear definition of interfaces, continuous integra-tion, utilization of unit testing, and a single set of doc-umentation and coding standards (Jenness et al. 2018).The primary implementation language is Python and,where necessary for performance reasons, C ++ .The LSST data management software has been proto-typed for over eight years. Besides processing simulated All components implemented in C ++ have been wrapped andexposed as Python modules to the rest of the system. Typicalusers should not have to work directly with the C ++ layer. Ivezi´c, Kahn, Tyson, Abel, Acosta, Allsman, Alonso, AlSayyad, Anderson, et al.
Figure 15.
A small region in the vicinity of globularcluster M2, taken from a coadd of SDSS Stripe 82 data pro-duced with LSST software stack prototypes. The coaddi-tion employs a novel “background matching” technique thatimproves background estimation and preserves the diffusestructures in the resulting coadd.
Figure 16.
A small portion, 4 (cid:48) × (cid:48) , of the HSC gri imagingof the COSMOS field. The limiting magnitude is about 27.5,roughly equivalent to 10-year LSST depth. LSST data ( § The LSST database design: Qserv
The scale of the LSST data release catalogs, in com-bination with desired targets for user concurrency andquery response times, present some engineering chal-lenges. The LSST project has been developing
Qserv ,a shared-nothing MPP (massively parallel processing)database system, to meet these needs (Wang et al. 2011;Becla & Wang 2014). Catalog data within Qserv is spa-tially partitioned, and hosted on shard servers runningon dedicated hardware resources within the LSST DataFacility. The shard servers locally leverage conventionalRDBMS (relational database management system) tech-nologies, running behind custom front-end codes whichhandle query analysis, rewrite, distribution, and resultaggregation. The Qserv shard servers also provide a fa-cility for cross-user synchronization of full-table scans inorder to provide predictable query response times whenserving many users concurrently. More details aboutQserv can be found in the LSST document LDM-135(Becla et al. 2017).2.7.
Simulating the LSST System
Throughout its design, construction and commission-ing, the LSST needs to be able to demonstrate that itcan achieve the requirements laid out in the Science Re-quirements Documents (SRD) given its design and as-delivered components, that the system can be calibrated he Large Synoptic Survey Telescope § § § ++ withthe overall framework and database interactions using P ython . The purpose of this design is to enable thegeneration of a wide range of data products for use bythe collaboration; from all-sky catalogs used in simula-tions of the LSST calibration pipeline, to studies of theimpact of survey cadence on recovering variability, tosimulated images of a single LSST focal plane.2.7.1.
The LSST Operations Simulator
The LSST Operations Simulator (Delgado et al. 2014)was developed to enable a detailed quantitative analysisof the various science tradeoffs described in this paper.It contains detailed models of site conditions, hardwareand software performance, and an algorithm for schedul-ing observations which will, eventually, drive the largelyrobotic observatory. Observing conditions include amodel for seeing derived from an extensive body of on-site MASS/DIMM (Multi-Aperture Scintillation Sensorand Differential Image Motion Monitor) measurementsobtained during site selection and characterization (seeFig. 1). It not only reproduces the observed seeing dis-tribution, but includes the auto-correlation spectrum ofseeing with time over intervals from minutes to seasons.Weather data are taken from ten years of hourly mea-surements at nearby Cerro Tololo. Thus the simulatorcorrectly represents the variation of limiting magnitudebetween pairs of observations used to detect NEOs and the correlation between, for example, seasonal weatherpatterns and observing conditions at any given point onthe sky. In addition, down time for observatory mainte-nance is also included.The signal-to-noise ratio of each observation is deter-mined using a sky background model which includes thedark sky brightness in each filter, the effects of seeingand atmospheric transparency, and a detailed model forscattered light from the Moon and/or twilight at eachobservation (Yoachim et al. 2016). The time taken tomove from one observation to the next is given by a de-tailed model of the camera, telescope, and dome. It in-cludes such effects as the acceleration/deceleration pro-files employed in moving the telescope, the dome, andthe wind screen, the time needed to damp vibrations ex-cited by each slew, cable wrap, the time taken for activeoptics lock and correction as a function of slew distance,and the time for filter changes and focal plane readout.Observations are scheduled by a ranking algorithm.After a given exposure, all possible next observationsare assigned a score which depends upon their locations,times, and filters according to a set of scientific require-ments which can vary with time and location. For ex-ample, if an ecliptic field has been observed in the r band, the score for another r -band observation of thesame field will initially be quite low, but it will rise intime to peak about an hour after the first observation,and decline thereafter. This algorithm results in obser-vations being acquired as pairs roughly an hour apart,which enables efficient association of NEO detections.To ensure uniform sky coverage, fields with fewer previ-ous observations will be scored more highly than thosewhich have already been observed more frequently.Once all possible next observations have been scoredfor scientific priority, their scores are modified accordingto observing conditions (e.g., seeing, airmass, and skybrightness) and to criteria such as slew time to movefrom the current position, time required to change fil-ters, etc. The highest-ranked observation is then per-formed, and the cycle repeats. The result of a simulatorrun is a detailed history of which locations on the skywere observed when, in what filter, and with what skybackground, seeing and other observing conditions. Ittakes a few days to produce a decade-long simulationusing an average PC.Results of the simulated surveys can be visualizedand analyzed using a Python-based package called theMetrics Analysis Framework (MAF; Jones et al. 2014).MAF provides tools to analyze the properties of a survey(e.g. the distribution of airmasses) through the creationof functions or metrics that are applied to OpSim out-puts. These metrics can express the expected technical2 Ivezi´c, Kahn, Tyson, Abel, Acosta, Allsman, Alonso, AlSayyad, Anderson, et al. performance of the survey, such as the number of visitsper field or the integrated depth after 10 years, as wellas the science capabilities or a survey, such as the num-ber of supernovae detected or the number of supernovaewith sufficient observations to have a well-characterizedlight curve. 2.7.2.
Catalog Generation
The simulated astronomical catalogs (CatSim; Con-nolly et al. 2014) are stored in an SQL database. Thisbase catalog is queried using sequences of observationsderived from the Operations Simulator. Each simulatedpointing provides a position and time of the observa-tion together with the appropriate sky conditions (e.g.,seeing, moon phase and angle, sky brightness and skytransparency). Positions of sources are propagated tothe time of observation (including proper motions forstars and orbits for Solar System sources). Magnitudesand source counts are derived using the atmospheric andfilter response functions appropriate for the airmass ofthe observation and after applying corrections for sourcevariability. The resulting catalogs are then formatted tobe output to users, or to be fed into an image simulator.The current version of the LSST simulation frameworkincorporates galaxies derived from an N-body simulationof a ΛCDM cosmology, quasars/AGNs, stars that matchthe observed stellar distributions within our Galaxy, as-teroids generated from simulations of our Solar Sys-tem, and a 3-D model for Galactic extinction. Stel-lar sources are based on the Galactic structure modelsof Juri´c et al. (2008) and include thin-disk, thick-disk,and halo star components. The distribution and col-ors of the stars match those observed by SDSS. Eachstar in the simulation is matched to a template spectralenergy distribution (SED). Kurucz (1993) model spec-tra are used to represent main-sequence F, G, and Kstars as well as RGB stars, blue horizontal branch stars,and RR Lyrae variables. SEDs for white dwarf starsare taken from Bergeron et al. (1995). SEDs for M,L, and T dwarfs are generated from a combination ofspectral models and stacks of spectra from the SDSS(e.g., Cushing et al. 2005; Bochanski et al. 2007; Burrowset al. 2006; Pettersen & Hawley 1989; Kowalski et al.2010). The adopted metallicity for each star is based ona model from Ivezi´c et al. (2008), and proper motionsare based on the kinematic model of Bond et al. (2010).Light curve templates are assigned to a subset of thestellar population so that variability may also be simu-lated. This assignment and variability are matched tovariability trends observed by the Kepler satellite, andaugmented by simulated distributions of RR-Lyrae andCepheids. For Galactic reddening, a value of E ( B − V ) is assigned to each star using the three-dimensional Galac-tic model of Amˆores & L´epine (2005). To provide con-sistency with the modeling of extragalactic fluxes in thesimulations, the dust model in the Milky Way integratedto 100 kpc is re-normalized to match the Schlegel et al.(1998) dust maps.Galaxy catalogs are derived from the Millennium sim-ulations of De Lucia et al. (2006). These models ex-tend pure dark matter N-body simulations to includegas cooling, star formation, supernovae and AGN, andare designed to reproduce the observed colors, luminosi-ties, and clustering of galaxies as a function of redshift.To generate the LSST simulated catalogs, a light cone,covering redshifts 0 < z <
6, was constructed from 58simulation snapshots 500 h − Mpc on a side. This lightcone extends to a depth of approximately r = 28 andcovers a 4.5 ◦ × ◦ footprint on the sky. Replicating thiscatalog across the sky simulates the full LSST footprint.As with the stellar catalog, an SED is fit to the colorsof each source using Bruzual & Charlot (2003) spectralsynthesis models. These fits are undertaken separatelyfor the bulge and disk components and, for the disk,include inclination-dependent reddening. Morphologiesare modeled using two S´ersic profiles. The bulge-to-diskratio and disk scale lengths are taken from De Luciaet al. (2006). Half-light radii for bulges are estimatedusing the empirical absolute-magnitude vs. half-lightradius relation given by Gonz´alez et al. (2009). Compar-isons between the redshift and number-magnitude dis-tributions of the simulated catalogs with those derivedfrom deep imaging and spectroscopic surveys showedthat the De Lucia et al. (2006) models under-predictthe density of sources at faint magnitudes and high red-shifts. To correct for these effects, sources are cloned inmagnitude and redshift space until their densities reflectthe average observed properties.Quasar/AGN catalogs are generated using the Bon-giorno et al. (2007) luminosity function for M B < − he Large Synoptic Survey Telescope OpenOrb software package (Granvik et al. 2009). With typically8000 sources per LSST field of view, this procedure sig-nificantly reduces the computational resources requiredto simulate asteroid ephemerides.2.7.3.
Image Simulations
The framework described above provides a parametrizedview of the sky above the atmosphere. Images are sim-ulated using two packages: GalSim (Rowe et al. 2015),and Phosim (Peterson et al. 2015). Galsim is a modu-lar and open-source package that provides a library forsimulating stars and galaxies through a range of modernastronomical telescopes. Point-spread-functions (PSFs)are treated as either analytic functions or modeled fromray-traced optics. Convolutions by the PSF can beapplied to parameterized galaxy profiles (e.g. S´ersicprofiles) or to directly observed images. Operations areapplied in Fourier space to enable an effective trade-offbetween speed of simulation and accuracy. GalSim iswritten in C++ with a Python API and is integratedwithin the LSST CatSim framework.Phosim is an open-source package that simulates im-ages by drawing photons from the spectral energy dis-tribution of each source (scaled to the appropriate fluxdensity based on the apparent magnitude of a source andaccounting for the spatial distribution of light for ex-tended sources). Each photon is ray-traced through theatmosphere, telescope and camera to generate a CCDimage. The atmosphere is modeled using a Taylor frozenscreen approximation (with the atmosphere describedby six layers). The density fluctuations within thesescreens are described by a Kolmogorov spectrum withan outer scale (typically 10 m to 200 m). All screensmove during an exposure, with velocities derived fromNOAA measurements of the wind velocities above theLSST site in Chile. Typical velocities are on the orderof 20 m s − , and are found to have a seasonable de-pendence that is modeled when generating the screens.Each photon’s trajectory is altered due to refraction asit passes through each screen.After the atmospheric refraction, the photons inPhoSim are reflected and refracted by the optical sur-faces within the telescope and camera. The mirrorsand lenses are simulated using geometric optics tech- niques in a fast ray-tracing algorithm and all opticalsurfaces include a spectrum of perturbations based ondesign tolerances. Each optic moves according to itssix degrees of freedom within tolerances specified bythe LSST system. Fast techniques for finding interceptson the aspheric surface and altering the trajectory ofa photon by reflection or wavelength-dependent refrac-tion have been implemented to optimize the efficiency ofthe simulated images. Wavelength and angle-dependenttransmission functions are incorporated within each ofthese techniques, including simulation of the telescopespider.Both GalSim and PhoSim model the propagation ofphotons through the silicon of the detector. The conver-sion probability, refraction as a function of wavelengthand temperature, and charge diffusion within the sili-con are modeled for all photons. Photons are pixelatedand the readout process simulated including blooming,charge saturation, charge transfer inefficiency, gain andoffsets, hot pixels and columns, the dependence of theimage size on intensity (a.k.a. the “brighter-fatter” ef-fect), and QE variations.An example of a simulated LSST image using PhoSimis shown in Fig. 17. ANTICIPATED DATA PRODUCTS AND THEIRCHARACTERISTICSThe LSST observing strategy is designed to maximizethe scientific throughput by minimizing slew and otherdowntime and by making appropriate choices of the fil-ter bands given the real-time weather conditions. Usingsimulated surveys produced with the Operations Sim-ulator described in § The Baseline LSST Surveys
The fundamental basis of the LSST concept is to scanthe sky deep, wide, and fast, and to obtain a dataset thatsimultaneously satisfies the majority of the science goals.We present here a specific realization, the so-called “uni-versal cadence”, which yields the main deep-wide-fastsurvey and meets our core science goals. However, atthis writing, there is a vigorous discussion of cadenceplans in the LSST community, exploring variants andalternatives that enhance various specific science pro-grams, while maintaining the science requirements de-scribed in the SRD.The main deep-wide-fast survey will use about 90%of the observing time. The remaining 10% of the ob-serving time will be used to obtain improved coverageof parameter space such as very deep ( r ∼
26) obser-vations, observations with very short revisit times ( ∼ Ivezi´c, Kahn, Tyson, Abel, Acosta, Allsman, Alonso, AlSayyad, Anderson, et al.
Figure 17.
A simulated image of a single LSST CCD using PhoSim (covering a 13 . × . region of the sky). Theimage is a color composite (Lupton et al. 2004) from a set of 30 second gri visits. minute), and observations of “special” regions such asthe Ecliptic plane, Galactic plane, and the Large andSmall Magellanic Clouds.3.1.1. The Main Deep-Wide-Fast Survey and itsextensions
The observing strategy for the main survey will beoptimized for the homogeneity of depth and number ofvisits. In times of good seeing and at low airmass, pref-erence is given to r -band and i -band observations. Asoften as possible, each field will be observed twice, with visits separated by 15–60 minutes. This strategy willprovide motion vectors to link detections of moving ob-jects in the Solar System, and fine-time sampling formeasuring short-period variability. The ranking criteriaalso ensure that the visits to each field are widely dis-tributed in position angle on the sky and rotation angleof the camera in order to minimize systematic effects ingalaxy shape determination.The universal cadence provides most of LSST’s powerfor detecting Near Earth Objects (NEO) and Kuiper he Large Synoptic Survey Telescope Figure 18.
The distribution of the r band visits on the skyfor a simulated realization of the baseline cadence. The skyis shown in the equal-area Mollweide projection in equato-rial coordinates (the vernal equinoctial point is in the center,and the right ascension is increasing from right to left). Thenumber of visits for a 10-year survey, normalized to the SRDdesign value of 184, is color-coded according to the legend.The three regions with smaller number of visits than themain survey (“mini-surveys”) are the Galactic plane (arc onthe right), the region around the South Celestial Pole (bot-tom), and the so-called “northern Ecliptic region” (upperleft; added in order to increase completeness for moving ob-jects). Deep drilling fields, with a much higher number ofvisits than the main survey, are also visible as small circles.The fields were dithered on sub-field scales and pixels withangular resolution of ∼
30 arcmin were used to evaluate anddisplay the coverage.
Belt Objects (KBOs) and naturally incorporates thesouthern half of the ecliptic within its 18,000 square de-grees, with a declination cut of about δ = +2 ◦ . Ad-ditional coverage of a crescent within 10 degrees of theNorthern ecliptic plane would sample the full azimuthaldistribution of KBOs, crucial for understanding the dif-ferent dynamical families in which they fall. Thus, weplan to extend the universal cadence to this region usingthe r and i filters only, along with more relaxed limitson airmass and seeing. Relaxed limits on airmass andseeing are also adopted for ∼
700 deg around the SouthCelestial Pole, allowing coverage of the Large and SmallMagellanic Clouds (Fig. 18).Finally, the universal cadence proposal excludes ob-servations at low Galactic latitudes, where the high stel-lar density leads to a confusion limit at much brightermagnitudes than those attained in the rest of the sur-vey. Within this region, the Galactic plane proposalprovides 30 observations in each of the six filters, dis-tributed roughly logarithmically in time (it may not benecessary to use the u and g filters for this heavily ex-tincted region). The resulting sky coverage for the LSST baseline ca-dence (known internally as minion 1016 ), based on de-tailed operations simulations, is shown for the r bandin Fig. 18. The anticipated total number of visits for aten-year LSST survey is about 2.45 million ( ∼ Mini-surveys and Deep Drilling Fields
Although the uniform treatment of the sky providedby the universal cadence proposal can satisfy the ma-jority of LSST scientific goals, roughly 10% of the timewill be allocated to other strategies that significantlyenhance the scientific return. These surveys aim to ex-tend the parameter space accessible to the main surveyby going deeper or by employing different time/filtersampling. We have already discussed three examplesof such mini-surveys: the Northern Ecliptic Spur to im-prove completeness of the asteroid and KBO population,the Southern Celestial Cap to extend the survey foot-print to the South Pole (thus providing coverage of theMagellanic Clouds), and the Galactic Plane survey toinclude low Galactic latitude fields.As an additional example of a mini-survey, consider aprogram that uses one hour of observing time per nightto observe a single pointing (9.6 deg ) to substantiallygreater depth in individual visits. Accounting for read-out time and filter changes, it could obtain about 50consecutive 15-second exposures in each of four filtersin an hour. If a field is visited every two days overfour months, about 600 deg can be observed with thiscadence over 10 years. Taking weather into account,the selected fields would each have on average about 40hour-long sequences of 200 exposures each. Each 15-second exposure in a sequence would have an equivalent5 σ depth of r ∼
24, and each filter subsequence whencoadded would be 2 magnitudes deeper than the mainsurvey visits ( r ∼ . Ivezi´c, Kahn, Tyson, Abel, Acosta, Allsman, Alonso, AlSayyad, Anderson, et al. main survey visits are coadded, they would extend thedepth to r ∼ | b | ∼
30 deg, they wouldinclude about 10 million stars with r <
21 observed withsignal-to-noise ratio above 100 in each visit. When sub-sequences from a given night were coadded, they wouldprovide dense time sampling to a faint limit of r ∼ . z ∼ .
2, with more densely sampled lightcurves than obtained from the universal cadence. Suchsequences would also serve as excellent tests of our pho-tometric calibration procedures.The LSST has already selected four distant extra-galactic survey fields that the project guarantees toobserve as Deep Drilling Fields with deeper coverage andmore frequent temporal sampling than provided by thestandard LSST observing pattern. These fields (EliasS1, XMM-LSS, Extended Chandra Deep Field-South,and COSMOS) are well-studied survey fields with sub-stantial existing multiwavelength coverage and otherpositive attributes. These four fields are only the firstchosen for deep-drilling observations. The project plansa community call for white papers suggesting additionaldeep drilling fields and other specialized observing ca-dences.3.2. Detailed Analysis of Simulated Surveys
As examples of analysis enabled by the OperationsSimulator ( § Expected Photometric Signal-to-Noise Ratio
The output of operations simulations is a data streamconsisting of a position on the sky and the time of obser-vation, together with observing conditions such as seeingand sky brightness. The expected photometric error in Table 2.
The Parameters From Eqs. 5 and 6 u g r i z ym skya θ b θ eff c γ d k m e C m f m C ∞ m h C m (2) i m a The expected median zenith sky brightness at CerroPach´on (AB mag arcsec − ). b The expected delivered median zenith seeing (FWHM, arc-sec). The seeing approximately scales with airmass, X , as X . . c The effective zenith seeing (arcsec) used for m computa-tion. d The band-dependent parameter from Eq. 5. e Adopted atmospheric extinction. f The band-dependent parameter from Eq. 6. g The typical 5 σ depth for point sources at zenith, assumingexposure time of 2 ×
15 sec, and observing conditions aslisted. For larger airmass the 5 σ depth is brighter; see thebottom row. h The loss of depth due to instrumental noise (assuming 9 e − per pixel and readout, and two readouts per visit). i Additive correction to C m when exposure time is doubledfrom its fiducial value to 60 sec. j The loss of depth at airmass of X = 1 . magnitudes (roughly the inverse of the signal-to-noiseratio) for a single visit can be written as σ = σ sys + σ rand , (4)where σ rand is the random photometric error and σ sys is the systematic photometric error (due to, e.g., im-perfect modeling of the point spread function, but notincluding uncertainties in the absolute photometric ze-ropoint). The calibration system and procedures aredesigned to maintain σ sys < .
005 mag. Based on SDSSexperience (Sesar et al. 2007), the random photometricerror for point sources, as a function of magnitude, is he Large Synoptic Survey Telescope by σ rand = (0 . − γ ) x + γ x (mag ) , (5)with x ≡ . m − m ) . Here m is the 5 σ depth (forpoint sources) in a given band, and γ depends on the skybrightness, readout noise, etc. Detailed determinationof the system throughput yields the values of γ listed inTable 2. The 5 σ depth for point sources is determinedfrom m = C m + 0 .
50 ( m sky −
21) + 2 . (0 . /θ eff ) ++1 .
25 log ( t vis / − k m ( X −
1) (6)where m sky is the sky brightness (AB mag arcsec − ), θ eff is the seeing (in arcsec), t vis is the exposure time(seconds), k is the atmospheric extinction coefficient,and X is airmass. Here the seeing corresponds to the“effective” seeing computed from the seeing FWHM fol-lowing the procedure described in Angeli et al. (2016).The seeing FWHM in each band is listed in the secondrow of Table 2, and the effective seeing is listed in thethird row of Table 2.The constants C m depend on the overall throughputof the instrument and are computed using our currentbest throughput estimates for optical elements and sen-sors. The resulting C m values are listed in Table 2 andin all six bands they imply single visit depths m (alsolisted in Table 2) that lie between the minimum and de-sign specification values from the Science RequirementsDocument listed in Table 1. The differences in perfor-mance between LSST and, for example, SDSS followdirectly from these relations .The structure of eq. 6 nicely illustrates decoupling be-tween the system sensitivity which is fully absorbed into C m and observing conditions specified by m sky , θ , t vis , k m and X . The computation of C m listed in Table 2assumed instrumental noise of 9 e − per pixel and perreadout, whose effect on m is significant only in the u band. This loss of depth due to instrumental noise,∆ C ∞ m , is listed in Table 2; it also corresponds to an addi-tive correction to C m when the exposure time t vis → ∞ . Eq. 5 can be derived from σ rand = N/S , where N is noise and S is signal, and by assuming that N = N o + αS . The constants N o and α can be expressed in terms of a single unknown constant γ by using the condition that σ rand = 0 . m = m . SDSS data typically reach a 5 σ depth for point sources of r = 22 . D = 2 .
22 m, an exposuretime of t vis = 54 sec, the median r band sky brightness of r sky =20 . − , the median seeing of θ = 1 . X = 1 .
3. In comparison, the LSST loses 0.32mag in depth due to shorter exposures, and gains 1.17 mag dueto larger aperture, 0.83 mag due to better seeing, and 0.20 magdue to fainter sky, for a net gain of ∼ To predict 5 σ depths for exposure time τ times longerthan the fiducial t vis = 30 sec., the following correctionshould be added to the values of C m listed in Table 2:∆ C m ( τ ) = ∆ C ∞ m − .
25 log (cid:20) (0 . C ∞ m ) − τ (cid:21) . (7)By definition, ∆ C m ( τ = 1) = 0. Again, this effect isonly substantial in the u band, as demonstrated by thevalues of ∆ C m ( τ = 2) listed in Table 2.The loss of depth at the airmass of X = 1 . The NEO Completeness Analysis
Detailed analyses of the LSST completeness for PHAsand NEOs are described in Jones et al. (2018), Vereˇs& Chesley (2017a,b), and Grav et al. (2016). After ac-counting for differences in their input assumptions andmodels, each of these independent works calculates acompleteness value which is consistent within a few per-cent. Here we briefly summarize the LSST project anal-ysis carried out in Jones et al. (2018); this approachis roughly the same for each of the studies mentionedabove.To assess the LSST completeness for PHAs, the PHApopulation is represented by a sample of orbits takenfrom the Solar System model of Grav et al. (2007).The simulated baseline survey is used to determinewhich PHAs are present in each exposure and at whatsignal-to-noise ratio they were observed. In additionto seeing, atmospheric transparency, and sky back-ground effects (see eq. 6), the signal-to-noise compu-tation takes into account losses due to non-optimal de-tection filters and object trailing. Using mean aster-oid reflectance spectra (DeMeo et al. 2009), combinedwith the LSST bandpasses, we calculate expected mag-nitudes and colors, assuming all PHAs are C type as-teroids, of V − m = ( − . , − . , . , . , . , . m = ( u, g, r, i, z, y ) to transform standard V bandmagnitudes to the magnitudes expected in each filter(Ivezi´c et al. 2001). Due to very red V − u colors, andthe relatively bright limiting magnitude in the y band,the smallest objects are preferentially detected in the griz bands. The correction for trailing is implementedby subtracting from the right-hand side of eq. 6∆ m trailing5 = 1 .
25 log (cid:0) . x (cid:1) (8) x = v t vis θ , (9)8 Ivezi´c, Kahn, Tyson, Abel, Acosta, Allsman, Alonso, AlSayyad, Anderson, et al. where the object’s velocity, v , is expressed in deg. day − .For the nominal exposure time of 30 seconds and θ = 0 . v = 0 .
25 deg. day − , typical for objects in the main as-teroid belt, and 0.46 mag for v = 1 . − , typicalof PHAs passing near Earth. PHAs are characterized bytheir “absolute magnitude” H , i.e., their apparent mag-nitude if they were placed 1 AU from both the Sun andthe Earth, with a phase angle of 0 ◦ . For a given albedo, H scales directly with diameter of the asteroid. ThePHA orbits are cloned over an H magnitude distribu-tion with dN/dH = 10 α H , with α = 0 .
33, in order toevaluate completeness as a function of H .An object is considered to be discovered if the objectwas detected on at least three nights within a windowof 15 days, with a minimum of two visits per night.The same criterion has been used in NASA studies,and is confirmed as reliable by a detailed analysis oforbital linking and orbit determination using the Mov-ing Object Processing System (MOPS) code (Vereˇs &Chesley 2017a,b; Jedicke et al. 2005) developed by thePan-STARRS project (and adopted by LSST in a col-laborative effort with Pan-STARRS). The MOPS soft-ware system and its algorithms are significantly moreadvanced than anything previously fielded for this pur-pose to date. Realistic MOPS simulations show > H ≤
22 (equivalent to D ≥
140 m) after 10 years of operations (Jones et al. 2018).This cadence spends 6% of the total observing time onNEO-optimized observations north of δ = +5 ◦ , andMOPS links objects with windows of 15 days. Thebaseline survey cumulative completeness as a functionof time for objects with H ≤
22 is shown in the up-per panel of Fig. 19, both with and without includingcontributions from current and on-going surveys. Thesefigures are likely to be uncertain at the level of ± H ≤
22 PHAs.By improving MOPS and increasing the MOPS linkingwindow from 15 to 30 days we can boost completenessby about 3%. By running the survey for an additionaltwo years, we can boost completeness by another 4%.Considering this ‘extended’ LSST in the context of ex-
Year . . . . . . N E O c o m p l e t e n e ss H NEOs
Existing resourcesLSST OnlyLSST + ExistingJPL NEO (
D > m) Year . . . . . . P H A c o m p l e t e n e ss H PHAs
Existing resourcesLSST OnlyLSST + Existing
LSST Baseline
Year . . . . . . N E O c o m p l e t e n e ss H NEOs
Existing resourcesLSST OnlyLSST + ExistingJPL NEO (
D > m) Year . . . . . . P H A c o m p l e t e n e ss H PHAs
Existing resourcesLSST OnlyLSST + Existing
Extended LSST Survey (12 years, 30 day linking windows)
Figure 19.
Cumulative completeness of the LSST surveyfor NEOs (left in each panel) and PHAs (right in each panel)brighter than a given absolute magnitude, H ≤
22 (related tothe size of the object and albedo; H =22 mag is equivalent toa typical 140 m asteroid). The top panel illustrates cumula-tive completeness for the LSST baseline cadence and MOPSconfiguration. In the baseline, LSST alone would discover66% of the PHAs with H ≤
22 (61% of NEOs); LSST com-bined with current and on-going surveys can discover 81%of PHAs (73% of NEOs). The bottom panel illustrates cu-mulative completeness when LSST is operated for 12 years,with extra visits around the ecliptic, and when the MOPSlinking window is increased to 30 days from the baseline 15.In this case, LSST alone could discover 74% of the PHAswith H ≤
22 (69% of NEOs); LSST combined with existingresources could discover 86% of PHAs (77% of NEOs). isting/ongoing surveys would result in a system-widecumulative completeness of 86% for PHAs (77% forNEOs), approaching the 90% required by the Congres-sional mandate (see lower panels of Fig. 19).3.2.3.
The Expected Accuracy of Trigonometric Parallaxand Proper Motion Measurements
To model the astrometric errors, we need to considerboth random and systematic effects. Random astro-metric errors per visit for a given star are modeled as θ/SN R , with θ = 700 mas and SN R determined usingeq. 6. Systematic errors of 10 mas are added in quadra-ture, and are assumed to be uncorrelated between differ-ent observations of a given object. Systematic and ran-dom errors become similar at about r = 22, and thereare about 100 stars per LSST sensor (0.05 deg ) to thisdepth (and fainter than the LSST saturation limit at r ∼
16) even at the Galactic poles. he Large Synoptic Survey Telescope Table 3.
The expected proper motion, par-allax and accuracy for a 10-year long baselinesurvey. r σ xy a σ π b σ µ c σ σ C e mag mas mas mas/yr mag mag21 11 0.6 0.2 0.01 0.00522 15 0.8 0.3 0.02 0.00523 31 1.3 0.5 0.04 0.00624 74 2.9 1.0 0.10 0.009 a Typical astrometric accuracy (rms per coor-dinate per visit). b Parallax accuracy for 10-year long survey. c Proper motion accuracy for 10-year longsurvey. d Photometric error for a single visit (two 15-second exposures). e Photometric error for coadded observations(see Table 1).
HSC data from the Subaru telescope reduced with theLSST software stack indicate that systematic errors of10 mas on spatial scales of several arcminutes are real-istic even at this stage of maturity of the code; resultsreported by DES (Bernstein et al. 2017) indicate as-trometric residuals of ∼ r <
22, which will provide exquisitecontrol of systematic astrometric errors as a function ofmagnitude, color and other parameters, and thus enableabsolute proper motion measurements.Given the observing sequence for each sky position inthe main survey as produced by the Operations Simula-tor ( § r ∼ z and y band detections. The impact on red stars issmaller due to a relatively small number of observationsin the u and g bands, but extremely red objects, suchas L and T dwarfs, will definitely have larger errors, de-pending on details of their spectral energy distributions.After the first three years of the survey, the proper mo-tion errors will be about five times as large, and parallaxerrors will be about twice as large, as the values given inTable 3; the errors scale as t − / and t − / , respectively.This error behavior is a strong independent argument fora survey lifetime of at least 10 years (c.f. § ∼ − per coordinate at r = 20 (Munn et al. 2004). Gaia isexpected to deliver parallax errors of 0.3 mas and propermotion errors of 0.2 mas yr − at its faint end at r ∼
20 (Perryman et al. 2001). Hence, LSST will smoothlyextend Gaia’s error vs. magnitude curve 4 magnitudesfainter (for illustration, see fig. 21 in Ivezi´c et al. 2012).3.3.
Data Products and Archive Services
Data collected by the LSST telescope and camera willbe automatically processed to data products – catalogs,alerts, and reduced images – by the LSST Data Manage-ment system ( § • Prompt products , designed to support thediscovery, characterization, and rapid follow-up oftime-dependent phenomena (“transient science”).These will be generated continuously every ob-serving night, by detecting and characterizingsources in images differenced against deep tem-plates. They will include alerts to objects that Historically, these have been referred to as “Level 1 DataProducts”, but going forward we prefer to use the more descrip-tive
Prompt Products designation. Note that the old terminologyis still in use in present-day LSST documents and code; new andupdated documents will gradually transition to the new, descrip-tive, nomenclature used in this paper. Ivezi´c, Kahn, Tyson, Abel, Acosta, Allsman, Alonso, AlSayyad, Anderson, et al. were newly discovered, or have changed bright-ness or position at a statistically significant level.The alerts to such events will be published within60 seconds of observation; we expect several mil-lion alerts per night.In addition to transient science, the prompt prod-ucts will support discovery and follow-up of ob-jects in the Solar System. Objects with motionssufficient to cause trailing in a single exposure willbe identified and flagged as such when the alertsare broadcast. Those that are not trailed will beidentified and linked based on their motion fromobservation to observation, over a period of a fewdays. Their orbits as derived by MOPS will bepublished within 24 hours of identification. Theefficiency of linking (and thus the completeness ofthe resulting orbit catalog) will depend on the fi-nal observing cadence chosen for LSST, as well asthe performance of the linking algorithm ( § • Data release products are designed to enablesystematics- and flux-limited science, and will bemade available in annual Data Releases . Thesewill include the (reduced and raw) single-epochimages, deep coadds of the observed sky, cata-logs of objects detected in LSST data, catalogs ofsources (the detections and measurements of ob-jects on individual visits), and catalogs of “forcedsources” (measurements of flux on individual vis-its at locations where objects were detected byLSST or other surveys). LSST data releases willalso include fully reprocessed prompt products,as well as all metadata and software necessaryfor the end-user to reproduce any aspect of LSSTdata release processing.A noteworthy aspect of LSST data release pro-cessing is that it will largely rely on multi-epochmodel fitting , or MultiFit , to perform near-optimal characterization of object properties.That is, while the coadds will be used to per-form object detection , the measurement of theirproperties will be performed by simultaneouslyfitting (PSF-convolved) models to all single-epochobservations. It is not yet clear to what extent wewill be able to make some of these measurements These have been referred to as “Level 2 Data Products” inthe past; as with their “Level 1” counterparts, we will use themore descriptive nomenclature going forward. The first-year data will probably be split into two data re-leases. on suitable linear combinations of input images(with careful propagation of PSFs and noise). Anextended source model – a constrained linear com-bination of two S´ersic profiles – and a point sourcemodel with proper motion – will generally be fit-ted to each detected object .Secondly, for the extended source model fits, theLSST will characterize and store the shape of theassociated likelihood surface (and the posterior),and not just the maximum likelihood values andcovariances. The characterization will be accom-plished by sampling, with up to ∼
200 (indepen-dent) likelihood samples retained for each object.For storage cost reasons, these samples may beretained only for those bands of greatest interestfor weak lensing studies.As described in § user-generated products at theLSST Data Facility, using the services offered within theLSST Science Platform ( § The LSST Science Platform
The LSST Science Platform (Juri´c et al. 2017a) repre-sents LSST’s vision for a large-scale astronomical dataarchive that can enable effective research with datasetsof LSST size and complexity. It builds on recent trendsin remote data analysis, and practical experiences in theastronomical context gathered by projects such as theJHU SciServer (Raddick et al. 2017), Gaia GAVIP (Vagget al. 2016), or NOAO Datalab (Fitzpatrick et al. 2016). For performance reasons, it is likely that only the point sourcemodel will be fitted in the most crowded regions of the Galacticplane. Formerly known as “Level 3 Data Products”. he Large Synoptic Survey Telescope aspects : • The web
Portal , designed to provide the essen-tial data access and visualization services througha simple-to-use website. It will enable queryingand browsing of the available datasets in ways theusers are accustomed to at archives such as IRSA,MAST, or the SDSS archive. • The
JupyterLab aspect, that will provide aJupyter Notebook-like interface and is geared to-wards enabling next-to-the-data remote analysis.A large suite of commonly used astronomical soft-ware, including the LSST software stack ( § • The
Web API aspect will expose the LSST DACservices to other software tools and services us-ing commonly accepted formats and protocols .This interface will open the possibility for remoteaccess and analysis of the LSST data set usingapplications that the users are already comfort-able with such as TOPCAT (Taylor 2005), or li-braries such as Astropy (Astropy Collaborationet al. 2013; Jenness et al. 2016). Furthermore, theoffered APIs will allow for federation with otherastronomical archives, bringing added value to theLSST dataset.Approximately 10% of the total budget for the LSSTData Facility compute and storage capacity has been re- http://jupyter.org/ For example, industry-standard protocols such as WebDAVmay be used to expose file data, or Virtual Observatory protocolssuch as TAP and SIAP may be used for access to catalogs andimages respectively. served for the LSST Science Platform needs, and to beshared by all LSST DAC users. Based on the currentplans and technology projections, these equate to ap-proximately 2,400 cores, 4 PB of file storage, and 3 PBof database storage at the beginning of LSST operations(in 2022). EXAMPLES OF LSST SCIENCE PROJECTSThe design and optimization of the LSST systemleverages its unique capability to scan a large sky areato a faint flux limit in a short amount of time. Themain product of the LSST system will be a multi-color ugrizy image of about half the sky to unprecedenteddepth ( r ∼ . ugriz images of about a quarter of the sky to r ∼ .
5, with twice as large seeing (see Figs. 20 and 21).A major advantage of LSST is the fact that this deep skymap will be produced by taking hundreds of shorter ex-posures (see Table 1). Each sky position within the mainsurvey area will be observed close to 1000 times, withtime scales spanning seven orders of magnitude (from30 sec to 10 years), and produce roughly thirty trillionphotometric measures of celestial sources.It is not possible to predict all the science that LSSTdata will enable. We now briefly discuss a few projectsto give a flavor of anticipated studies, organized by thefour science themes that drive the LSST design (al-though some projects span more than one theme). Foran in-depth discussion of LSST science cases, we referthe reader to the LSST Science Book, and more special-ized documents discussing cosmology (LSST Dark En-ergy Science Collaboration 2012; Zhan & Tyson 2018),galaxy science (Robertson et al. 2017), and synergy withother ground-based and space-based facilities (Najitaet al. 2016; Jain et al. 2015; Rhodes et al. 2017).4.1.
Probing Dark Energy and Dark Matter
A unique aspect of LSST as a probe of dark energyand dark matter is the use of multiple cross-checkingprobes that reach unprecedented precision (see Fig. 22).Any given probe constrains degenerate combinations ofcosmological parameters, and each probe is affected bydifferent systematics, thus the combination of probes al-lows systematics to be calibrated and for degeneraciesto be broken. Dark energy manifests itself in two ways.The first is the relationship between redshift and dis-tance (the Hubble diagram), or equivalently the expan-sion rate of the Universe as a function of cosmic time.The second is the rate at which matter clusters withtime. Structure formation involves a balance betweengravitational attraction of matter over-densities and the2
Ivezi´c, Kahn, Tyson, Abel, Acosta, Allsman, Alonso, AlSayyad, Anderson, et al.
Figure 20.
A comparison of ∼ . × . images ofthe same area of sky (centered on α =9 h (cid:48) (cid:48)(cid:48) and δ =30 ◦ (cid:48) (cid:48)(cid:48) ) obtained by the SDSS (top, r < .
5) and the DeepLens Survey (bottom, r < . ∼ gri color images tothe same depth ( ∼
200 for the riz composites) of each pointover half the Celestial sphere (18,000 deg , equivalent to 1.15million ∼ . × . regions), and with better seeing.After their coaddition, the final image will be another ∼ r = 27 . Figure 21.
A comparison of angular resolution for 20 × images obtained by the SDSS (top, median seeingof 1.5 arcsec) and expected from LSST (bottom, seeing of0.7 arcsec). The images show a lensed SDSS quasar (SDSSJ1332+0347, Morokuma et al. 2007); the bottom image wastaken with Suprime-cam at Subaru. Adapted from Bland-ford & LSST Strong Lensing Science Collaboration (2007). rapid expansion of the background. Thus, quantifyingthe rate of growth of structures from early times untilthe present provides additional tests of the energy con-tents of the Universe and their interactions. he Large Synoptic Survey Telescope w = w + w a (1 − a )) is a poor representation of moregeneral dark energy theories. Barnard et al. (2008)showed that in a high-dimensional dark energy modelspace, LSST data could lead to a hundred- to thousand-fold increase in precision over its precursor experiments,thereby confirming its status as a premier Stage IV ex-periment in the sense of Albrecht et al. (2006).The power and accuracy of LSST dark energy anddark matter probes are a result of the enormous sam-ples that LSST will have, including several billion galax-ies and millions of Type Ia supernovae. At i < . >
20 for point sources), the photometry of galax-ies will be of high enough quality to provide photomet-ric redshifts with an RMS accuracy ( σ/ (1 + z )) of 2%over the range 0 . < z < . σ/ (1 + z ) to ∼ i = 25 . − with shapes measured wellenough for weak lensing measurements (Chang et al.2013, 2015), with the number realized in practice beingdependent on the performance of the deblending andshape measurement algorithms. The median redshiftfor this sample will be z ∼ z ∼
2. It will be possible to further improve photometricredshift calibration by cross-correlating the photometricsample with redshift surveys of galaxies and quasars in
Figure 22.
Constraints on the dark energy equation ofstate ( w = w + w a (1 − a )) from LSST cosmological probes.The various ellipses assume constraints from BAO (dashedline), cluster counting (dash-dotted line), supernovae (dot-ted line), WL (solid line), joint BAO and WL (green shadedarea), and all probes combined (yellow shaded area). TheBAO and WL results are based on galaxy–galaxy, galaxy–shear, and shear–shear power spectra only. Adding otherprobes such as strong lensing time delay and higher-ordergalaxy and shear statistics will further improve the con-straints. While the details of the contours will change slightlyas the survey parameters are updated, the key point remainsthat this combination of dark energy probes results in con-tours with different degeneracy directions, and hence theircombination results in tight constraints on the dark energyequation of state. the same fields (Newman 2008; Matthews & Newman2010; M´enard et al. 2013; Davis et al. 2017).The main LSST observables in the context of darkenergy and matter are described below. • The joint analysis of shear–shear, galaxy–shear,and galaxy–galaxy correlation functions has be-come standard in analyses of precursor datasets(e.g. DES Collaboration et al. 2017; Joudaki et al.2018). WL and LSS are highly complementaryprobes, and the combination of their auto- andcross-correlations will constrain the properties ofthe late-time accelerated expansion while provid-ing internal cross-checks for marginalizing oversystematic uncertainties (e.g., Mandelbaum 2017).These measurements consist of the two-point auto-and cross-correlations of shear and positions forbillions of galaxies across ∼
10 redshift bins. Asdescribed in the following two items, the galaxy-4
Ivezi´c, Kahn, Tyson, Abel, Acosta, Allsman, Alonso, AlSayyad, Anderson, et al.
Figure 23.
Marginalized 1 σ errors on the comoving distance(open triangles) and growth factor (open circles) parametersfrom the joint analysis of LSST LSS and WL (galaxy–galaxy,galaxy–shear, and shear–shear power spectra) with a conser-vative level of systematic uncertainties in the photometricredshift error distribution and additive and multiplicativeerrors in the shear and galaxy power spectra. The maximummultipole used for WL is 2000, and that for LSS is 3000[with the additional requirement that ∆ δ ( (cid:96)/D A ; z ) < . z ) be-tween z = 0 and 5, and the distance parameters start at z = 0 .
14. The error of each distance (growth) parameter ismarginalized over all the other parameters including growth(distance) parameters. The joint constraints on distance arerelatively insensitive to the assumed systematics (Zhan et al.2009). galaxy and galaxy-shear correlation functions pro-vide additional probes of dark energy and darkmatter. • The galaxy–shear correlation function probes thegrowth of dark matter large-scale structure andis a diagnostic of the underlying cosmology. Thecombination with the galaxy–velocity correlationfunction estimated from currently planned spec-troscopic surveys could test General Relativity andits variants at high redshift (Reyes et al. 2010). • The galaxy–galaxy correlation function is vital toconstrain the galaxy bias impacting the galaxy-shear correlation and is therefore an essential com-ponent in the joint analysis of LSS and WL. Inaddition, the presence of Baryon Acoustic Oscilla-tions in the galaxy angular correlation functions isa strong cosmological probe on its own. The sound horizon at decoupling, which is imprinted on themass distribution at all redshifts and calibratedwith the CMB, provides a standard ruler to mea-sure the angular diameter distance as a function ofredshift (Eisenstein et al. 1998; Cooray et al. 2001;Blake & Glazebrook 2003; Hu & Haiman 2003;Linder 2003; Seo & Eisenstein 2003). LSST photo- z BAO will achieve percent-level precision on theangular diameter distance at ∼
10 redshifts loga-rithmically spaced between z = 0 . k to within 0.001. • Higher-order shear and galaxy statistics and shearpeak counts can improve dark energy constraintsand provide self-calibration of various systematics(Takada & Jain 2004; Dolney et al. 2006; Hutereret al. 2006; Petri et al. 2016). They are also probesof both primordial non-Gaussianities and thosecaused by non-linear structure. • Primordial non-Gaussianity is also probed by thelarge-scale power of any biased tracer of the matteroverdensities (Dalal et al. 2008). Although mea-surements of the galaxy power spectrum on verylarge scales are challenging due to sky systemat-ics (Leistedt et al. 2014) and cosmic variance, theprospect of using multiple tracers of the same fieldcould significantly improve the constraining powerfor this observable (Seljak 2009). Similar measure-ments of the large-scale power will also be used totest phenomenological models of clustering darkenergy (Takada 2006). • Similarly, weak lensing magnification tomography(Morrison et al. 2012) offers a complementaryprobe of a mix of cosmic geometry and growthof dark matter structure. • The two LSST observing programs are comple-mentary in the supernova samples they will pro-vide. The main survey will obtain light curvesin six bands and photometric redshifts of about400,000 photometrically-classified Type Ia super-novae that can be used for cosmological distancemeasurements, with further spectroscopic follow-up of a sub-sample of their host galaxies. Such asample will not only provide larger statistics forthe study of the Type Ia population in the uni-verse, but will also be spread across the full 18,000 he Large Synoptic Survey Telescope LSST main survey footprint, allowing dif-ferent probes of the large scale structure of thelow redshift universe. This sample of supernovaecan be used as a tracer of large scale structureby directly probing the gravitational potential ofstructure through inferences of their peculiar ve-locities (Gordon et al. 2007; Bhattacharya et al.2011; Howlett et al. 2017), weak lensing of super-nova brightnesses (Dodelson & Vallinotto 2006;Quartin et al. 2014; Macaulay et al. 2017; Scov-acricchi et al. 2017), and the local bulk flow (Riess2000; Dai et al. 2011; Turnbull et al. 2012; Feindtet al. 2013; Huterer et al. 2015), as well as low red-shift constraints on the isotropy of the universe(Antoniou & Perivolaropoulos 2010; Colin et al.2011; Campanelli et al. 2011; Cai et al. 2013; Ja-vanmardi et al. 2015). The rapidly sampled deepdrilling fields, possibly coadded over short timescales, will yield well-sampled light curves of tensof thousands of supernovae to redshifts peakingaround z ∼ . • Cosmological analyses can be carried out using SN,WL, and LSS in subsets of the LSST data in differ-ent regions of the sky, testing fundamental cosmo-logical assumptions of homogeneity and isotropy(e.g., Zhan et al. 2009). • The shape of the power spectrum of dark matterfluctuations measured by LSST weak lensing willconstrain the sum of neutrino masses with an ac-curacy of 0.04 eV or better (Cooray 1999; Song& Knox 2004; Hannestad et al. 2006). Given thecurrent constraints on neutrino mass mixing, thisis at the level to determine whether there is aninverted neutrino mass hierarchy, a fundamentalquestion in particle physics. • Tens of thousands of galaxy-galaxy lenses will pro-vide the needed statistics to probe dark matterhalo profiles and substructure (e.g., Mandelbaumet al. 2006; Vegetti et al. 2012). The image fluxesin several thousand well-measured strongly lensedquasars will enable constraints of the dark matter mass function on small scales (Dalal & Kochanek2002). • The abundance of galaxy clusters as a function ofmass and redshift is sensitive to cosmological pa-rameters (SciBook, Ch. 13; von der Linden et al.2014). LSST will produce a large catalog of clus-ters detected through their member galaxy popu-lation to redshift z ∼ .
2. In addition, LSST willidentify optical counterparts and provide deep op-tical imaging for clusters detected in other wave-bands (e.g., Staniszewski et al. 2009). • The clustering properties of those same galaxyclusters will also be used to constrain cosmologi-cal parameters (Mo et al. 1996; Mana et al. 2013),to marginalize over uncertainties in the mass-observable relation and photometric redshift un-certainties (Oguri & Takada 2011), and to con-strain the effects of super-sample covariance inthe two-point functions of WL and LSS (Hu &Kravtsov 2003; Takada & Spergel 2014). • LSST will discover several hundred galaxy clustersthat produce multiple-image lenses of backgroundobjects. Cluster mass reconstruction based on themultiple image positions can probe the cluster in-ner mass profile, and can provide a separate test ofcosmology, especially in cases with strongly lensedbackground objects at different redshift (Porciani& Madau 2000; Oguri & Kawano 2003). • Time delays of galaxy-scale lensed quasars will al-low one to measure Hubble’s constant (e.g., Suyuet al. 2010; Bonvin et al. 2017) in hundreds of sys-tems; sub-percent level precision in H ( z ) should beachievable (Coe & Moustakas 2009; Treu & Mar-shall 2016), providing a further independent darkenergy probe. LSST will also discover between500 and 1000 strongly lensed Type Ia supernovae(Goldstein & Nugent 2017; Goldstein et al. 2017),which will provide hundreds of additional high-quality time delays. Time delays for quasars mul-tiply lensed by clusters as a function of redshift arean independent test of dark energy (Kundi´c et al.1997). The natural timescale (many months toyears) is well matched to the LSST survey (Oguri& Marshall 2010). • Standard sirens are a new cosmological probe,demonstrated by the recent discovery of a binaryneutron star merger by LIGO with an electromag-netic counterpart (Abbott et al. 2017a), which wasused to constrain the Hubble parameter to roughly6
Ivezi´c, Kahn, Tyson, Abel, Acosta, Allsman, Alonso, AlSayyad, Anderson, et al.
15% precision (Abbott et al. 2017b). Constraintsfrom standard sirens are independent of the localdistance ladder, with the primary uncertainties be-ing the local velocity field and the inclination an-gle of the system. Scolnic et al. (2018) estimate oforder 70 such systems could be found with LSST.4.2.
Taking an Inventory of the Solar System
The small bodies of the Solar System, such as main-belt asteroids, the Trojan populations of the giant plan-ets and the Kuiper Belt objects, offer a unique insightinto its early stages because they provide samples ofthe original solid materials of the solar nebula. Under-standing these populations, both physically and in theirnumber and size distribution, is a key element in testingvarious theories of Solar System formation and evolu-tion.The baseline LSST cadence will result in orbital pa-rameters for several million objects; these will be dom-inated by main-belt asteroids, with light curves andmulti-color photometry for a substantial fraction of de-tected objects. The LSST sample of asteroids with ac-curate orbits and multi-color light curves will be 10 to100 larger than currently available sample. LSST willmake a significant contribution to the Congressional tar-get completeness of 90% for PHAs larger than 140 m( § r ∼ . in situ exploration far beyond the edge ofthe Kuiper belt at ∼
50 AU. Because most of these ob-jects will be observed several hundred times, accurateorbital elements, colors, and variability information willalso be available.The following are some examples of the LSST scienceopportunities in Solar System science: • Studies of the distribution of orbital elements forover 5 million main-belt asteroids as a functionof color-based taxonomy (see Fig. 24) and size;size distributions of asteroid families (Parker et al.2008) and their correlations with age (Jedicke et al.2004; Nesvorn´y et al. 2005); dynamical effects(Bottke et al. 2001); and studies of object shapesand spin states using light curve inversion tech-niques (Pravec & Harris 2000; Durech et al. 2009). • Studies of transient mass loss in asteroids (ac-tive asteroids or main belt comets, Hsieh & Je-witt 2006); such objects will appear extended inthe sensitive LSST images. Only a few such ob-jects are currently known (Jewitt et al. 2011; Je-witt 2012); LSST will increase the sample of suchobjects to ∼ Figure 24.
An example of color-based asteroid taxonomy.The figure shows the distribution of asteroids in the propersemi-major axis vs. sin( i ) plane for 45,000 asteroids with col-ors measured by SDSS (Parker et al. 2008). The color of eachdot is representative of the object’s color. Note the strongcorrelation between asteroid families (objects with similarorbital elements) and colors. There are at least five differenttaxonomic types distinguishable with SDSS measurements;LSST color measurements of asteroids will be more thantwice as accurate and will increase the number of objectsby roughly two orders of magnitude. • Studies of the distribution of orbital elements forabout 100,000 NEOs as a function of color and size(Rabinowitz 1993; Dandy et al. 2003); correlationswith the analogous distributions for main-belt ob-jects, and studies of object shapes and structureusing light curves. • Studies of the distribution of orbital elements forclose to 300,000 Jovian Trojan asteroids as a func-tion of color and size (Jewitt et al. 2000; Yoshida& Nakamura 2005; Szab´o et al. 2007); the searchfor dynamical families (Knezevic & Milani 2005);studies of shapes and structure using light curves. • Studies of the distribution of orbital elements forabout 30,000 TNOs (see Fig. 25) as a function ofcolor and size; the search for dynamical families(Marcus et al. 2011); studies of shapes and struc-ture using light curves (Duncan et al. 1995; Tru-jillo et al. 2001; Gladman et al. 2001; Bernsteinet al. 2004; Elliot et al. 2005; Jones et al. 2006;Doressoundiram et al. 2007). • An unbiased and complete census of both Jupiter-family and Oort-cloud comets. These comets he Large Synoptic Survey Telescope Figure 25.
The LSST detection limits for distant Solar Sys-tem objects as a function of distance. Moving objects withdiameters as small as 100 m in the main asteroid belt and100 km in the Kuiper Belt (TNOs) will be detected in indi-vidual visits, depending on the albedo. Specialized deeperobservations (see § will have detailed six-band high-resolution imagesextending to low surface brightness, in multiplepoints through their orbits, allowing detailed stud-ies of activity as a function of distance from theSun (Lowry et al. 1999; A’Hearn 2004). LSSTwill discover an unprecedentedly large number ofcomets with typically 50 observations per objectspread throughout their orbits during the 10-yearsurvey, and will help us to constrain models ofthe origin of comets (Solontoi 2010; Silsbee &Tremaine 2016). Combining the CN productionrates determined from observations in the u band-pass, as a proxy for overall gas activity, with thenon-volatile production rate calculated from thecontinuum-sensitive r , i , and z bands allows forthe determination of the gas-to-dust ratio. The re-lationship between the gas-to-dust ratio in cometsand their dynamical class (and places of forma-tion) is a fundamental, and still unresolved, ques-tion in cometary science (see e.g., A’Hearn et al.1995; Bockel´ee-Morvan & Biver 2017). • Searching for objects with perihelia out to severalhundred AU. For example, an object like Sedna(Brown et al. 2004) would be detectable at 130 AU.This will result in a much larger, well-understoodsample of inner Oort Cloud objects like Sedna and 2012 VP113 (Trujillo & Sheppard 2014). Study-ing the distribution of their orbits (in particularincluding any clustering in the argument of peri-helion) will test models predicting the existence ofa planetary-mass object beyond Neptune, a pro-posed Planet 9 (Trujillo & Sheppard 2014; Batygin& Brown 2016; Brown & Batygin 2016; Sheppard& Trujillo 2016; Brown 2017). Depending on theproposed Planet 9’s on-sky location and bright-ness, it may be possible for LSST to directly detectit in the wide survey images (Batygin & Brown2016; Brown & Batygin 2016; Sheppard & Trujillo2016; Brown 2017). • Mapping the propagation of coronal mass ejectionsthrough the Solar System using induced activity ina large sample of comets at different heliocentricdistances (SciBook Ch. 5). • Probing the inventory and frequency of interstel-lar asteroids/comets. The recent Pan-STARRS1discovery of the interstellar object 1I/2017 U1(‘Oumuamua) (Bacci et al. 2017) has shown thepower of large, complete all-sky surveys to unearthrare and exciting classes of objects. LSST will besome three magnitudes more sensitive than cur-rent NEO surveys (like Pan-STARRS1), and willcover more sky more often. Therefore, LSST islikely to find more interstellar objects, and morefrequently. Estimates from Cook et al. (2016), En-gelhardt et al. (2017), and Trilling et al. (2017)suggest that LSST will increase the number ofsuch rare objects by an order of magnitude which,among other outcomes, will help constrain the fre-quency and properties of planetary system forma-tion in the solar neighborhood.4.3.
Exploring the Transient Optical Sky
Time domain science will greatly benefit from LSST’sunique capability to simultaneously provide large areacoverage, dense temporal coverage, accurate color infor-mation, good image quality, and rapid data reductionand classification. Since LSST extends time-volume-color space 50-100 times over current surveys (e.g., Djor-govski et al. 2013) it will facilitate new population andstatistical studies and also the discovery of new classes ofobjects. LSST data products will enable many projectsincluding: • Discovery and characterization of thousands of hotJupiters in exoplanetary systems via the transitmethod (Wright et al. 2012). LSST will extendthe extrasolar planet census to larger distances8
Ivezi´c, Kahn, Tyson, Abel, Acosta, Allsman, Alonso, AlSayyad, Anderson, et al.
Figure 26.
The phase space of cosmic explosive and erup-tive transients as represented by their absolute V band peakbrightness and the event timescale, defined as the time takento drop one magnitude in V band brightness from peak lu-minosity (adapted from Kulkarni et al. (2007) and Kasliwal(2011)). The locus of the Classical Novae is as described indella Valle & Livio (1995). LSST will open up large regionsof this phase space for systematic exploration by extendingtime-volume space more than 100 times over existing sur-veys. within the Galaxy, thus enabling detailed studiesof planet frequency as a function of stellar metal-licity and parent population (e.g., Hartman et al.2009; Bayliss & Sackett 2011). The out-of-transitvariability of exoplanet host stars will also providecharacterization of the system via flaring behav-ior and stellar age via gyrochronology, the latterhelping constrain theories of tidal evolution andmigration in giant planets. • Gravitational microlensing in the Milky Way (seeHan 2008) as well as in the Local Group and be-yond (de Jong et al. 2008). • Studies of dwarf novae, including their use asprobes of stellar populations and structure in theLocal Group (Neill & Shara 2005; Shara 2006;Shen & Bildsten 2009). Population studies of theend points of binary evolution, mapping the distri-bution and quantifying the demographics of longand short orbital period dwarf novae, and distin-guishing recurrent from normal novae. Regularcadence, long term color observation on a largesample of galactic sources will enable the identi-fication of CVs containing highly magnetic whitedwarfs, that are red due to cyclotron emission from the magnetic accretion column and in the low statefor the majority of the 10-year survey. • Studies of transients from poorly-constrainedstages of stellar evolution including stellar erup-tions, luminous blue variable (LBV), stellar merg-ers, and helium core flashes leading to white dwarfformation. We will be able to identify the progeni-tors of eruptive transients in the deep LSST stacksand even look for faint precursor eruptions. Wewill also constrain the rates of individual erup-tion subclasses (Smith 2014) by detecting them ingalaxies out to tens of Mpc. • A census of light echoes of historical explosive anderuptive transients in the Milky Way and LocalGroup through high resolution time series. • Studies of known and unusual SN populations andparameterization of their light curves (e.g., H¨oflichet al. 1998; Wang et al. 2003; Howell et al. 2007;Kowalski et al. 2008; Hicken et al. 2009; Foley2012; Bianco et al. 2014; Arcavi et al. 2017), in-cluding late-time observations of rapidly-evolvingtransients to deep limits, critical for ascertainingtheir nature. Measurements of intrinsic rates forboth peculiar transients (e.g., Drout et al. 2014)and for SN as a function of sub-type and host envi-ronment properties (e.g., metallicity; Graur et al.2017). • A deep search for new populations of novae andsupernova progenitors (Smartt 2009; Thompsonet al. 2009; Smith et al. 2011, see Fig.26) boththrough direct imaging and through the detectionof SN precursor events (Ofek et al. 2013), charac-terization of pre-SN variability of SN progenitorsand the frequency of pre-SN outbursts. • The discovery of strongly lensed SNe; 500 − • A large, well characterized sample of super lumi-nous supernovae including object at redshift ashigh as z = 2 .
5, a sample large enough to be he Large Synoptic Survey Telescope w and Ω m (Scovacricchi et al. 2016). • Studies of optical bursters (those varying fasterthan 1 mag hr − ) to r ∼
25 mag. • Detection and measurement of gamma-ray burstafterglows and transients (e.g., Zhang & M´esz´aros2004; Zhang et al. 2006; Kann et al. 2010) to highredshift ( ∼ • Large scale studies of stellar tidal disruptions bynuclear supermassive black holes (e.g., Evans &Kochanek 1989; Gezari et al. 2008; Strubbe &Quataert 2009; Bloom et al. 2011; Gezari 2012;Komossa 2015), as well as binary supermassiveblack holes in the in-spiral phase (e.g., Cuadraet al. 2009; Coughlin et al. 2017a). Persistent ob-servations leading to complete lightcurves (otherthan the seasonal gaps) of long duration eventslike TDEs. Measurements of rates as function ofgalaxy type, redshift, and level of nuclear activity.An assessment of the diversity of these events interms of total power, effective temperature, andjet launching efficiency. • A study of quasar variability using accurate, mul-ticolor light curves for a few million quasars, lead-ing to constraints on the accretion physics andnuclear environments (de Vries et al. 2003; Van-den Berk et al. 2004; MacLeod et al. 2010; Jianget al. 2017). Relations between quasar variabil-ity properties and luminosity, redshift, rest-framewavelength, time scale, color, radio-jet emission,black-hole mass, and Eddington-normalized lumi-nosity will be defined with massive statistics, in-cluding the potential to detect rare but importantevents such as jet flares and obscuration events.Microlensing events will also be monitored in the ∼ • The superb continuum light curves of AGN willenable economical “piggyback” reverberation-mapping efforts using spectroscopy of emissionlines (e.g., Chelouche & Daniel 2012; Shen et al.2015; Grier et al. 2017). These results willgreatly broaden the luminosity-redshift plane ofreverberation-mapped AGNs with black-hole massestimates. For LSST data alone, the inter-bandcontinuum lags will provide useful structural in-formation. • Optical identification of transients and variablesdetected in other electromagnetic wavebands,from gamma rays to radio. Examples includeoptical and gamma ray variability in blazars (Ho-vatta et al. 2014), radio transients associated withtidal disruption flares (Giannios & Metzger 2011),and radio counterparts to supernovae and GRBs(Gal-Yam et al. 2006). Deep optical observationswith LSST may also help illuminate the natureof fast radio bursts (FRBs, Lorimer et al. 2007;Thornton et al. 2013). • Optical identification of counterparts to non-electromagnetic sources, such as gravitationalwaves (GW) and neutrino events (LIGO , ICE-CUBE ). LSST’s unique ability to characterizethe faint variable sky over large areas will beimportant for the detection of GW associatedsources, with an estimate of ∼ experiment hasled to the discovery of four GW events in lessthan a year. The binary neutron star mergerevent GW170817 was accompanied by emissiondetected across the entire electromagnetic spec-trum (Abbott et al. 2017a). The optical/NIRemission had two distinct components, a blueemission (which peaked and then faded away on atime scale of a few days) and a redder componentthat persisted for ∼
15 days. This longer-lastingcomponent arose from the radioactive decay ofheavy elements synthesized during the NS merger,a “kilonova” (AT 2017gfo). While both thesecomponents had been predicted (Metzger 2017),the ∼
100 kilonova sample that LSST is expectedto generate will enable comparative studies ofthese transients, allowing us to understand howthe presence and relative luminosity of the twocomponents varies to the properties of the binarysystem (e.g., mass) and its remnant. Furthermore,LSST will be important for identifying the opticaltransient corresponding to LIGO events in the firstplace, eliminating false positives (Nissanke et al.2013; Metzger & Berger 2012; Cowperthwaite &Berger 2015; Coughlin et al. 2017b). At 24th mag,rejecting thousands of false positives from othernew transients appearing during the imaging of http://icecube.wisc.edu http://public.virgo-gw.eu/language/en/ Ivezi´c, Kahn, Tyson, Abel, Acosta, Allsman, Alonso, AlSayyad, Anderson, et al.
Figure 27.
The g − r vs. u − g color-color diagram forabout a million point sources from the SDSS Stripe 82 area.Accurate multi-color photometry contains information thatcan be used for source classification and determination ofdetailed stellar properties such as effective temperature andmetallicity. LSST will enable such measurements for billionsof stars. the GW event area requires a strategy of multiplepasses in different filters.4.4. Mapping the Milky Way
The LSST will map the Galaxy in unprecedented de-tail, and by doing so revolutionize the fields of GalacticAstronomy and Near-field Cosmology. The great detailwith which the Milky Way can be studied complementsthe statistical power of extra-galactic observations. Theoverarching goal of near-field cosmology is to use spa-tial, kinematic, and chemical measurements of stars toreveal the structure and evolution history of the MilkyWay and its environment. LSST will reveal this fossilrecord in great detail and provide a Rosetta Stone forextragalactic astronomy by setting the context withinwhich we interpret these broader datasets. Moreover,different candidate supersymmetric particle dark mattermodels predict different mass clustering on small scales,and thus different mass functions for subhalos of theMilky Way. Thus the LSST census of faint satellitesand stellar streams in the halo will offer a unique meansto constrain the particle nature of dark matter.The LSST will produce a massive and exquisitely ac-curate photometric and astrometric dataset for about20 billion Milky Way stars. The coverage of the Galac-tic plane will yield data for numerous star-forming re-gions, and the y band data will penetrate through theinterstellar dust layer. Photometric metallicity measure- ments (see Figs. 27 and 28) will be available for about200 million main-sequence F/G stars which will samplethe halo to distances of 100 kpc (Ivezi´c et al. 2008; Anet al. 2013) over a solid angle of roughly 20,000 deg .No other existing or planned survey will provide such apowerful dataset to study the outer halo: Gaia is fluxlimited at r = 20, and the Dark Energy Survey (Ros-setto et al. 2011) and Pan-STARRS both lack observa-tions in the u band, necessary for estimating metallicity.The LSST in its standard surveying mode will be ableto detect RR Lyrae variables (pulsating stars and stan-dard candles) and classical novae (exploding stars andstandard candles) at a distance of 400 kpc and henceexplore the extent and structure of our halo out to halfthe distance to the Andromeda galaxy. Thus, the LSSTwill enable studies of the distribution of main-sequencestars beyond the presumed edge of the Galaxy’s halo(see Fig. 29), of their metallicity distribution through-out most of the halo, and of their kinematics beyondthe thick disk/halo boundary. LSST will also obtain di-rect distance measurements via trigonometric parallaxbelow the hydrogen-burning limit for a representativethin-disk sample.In addition to the study of hydrogen burning stars,LSST will uncover the largest sample of stellar rem-nants to date. Over 97% of all stars eventually exhausttheir fuel and cool to become white dwarfs. Given theage of the Galactic halo, a significant fraction of themass in this component may reside in these remnantstars (e.g., Alcock et al. 2000; Tisserand et al. 2007) andtherefore their discovery directly constrains the Galacticmass budget. These large populations of disk and halowhite dwarfs will provide unprecedented constraints onthe luminosity function of these stars, which will di-rectly yield independent ages for the Galactic disk andhalo (e.g., through the initial-final mass relation, Kaliraiet al. (2008)).The sky coverage of LSST naturally targets both fieldstars and star clusters. To date, no systematic surveyof the stellar populations of Southern hemisphere clus-ters has been performed (e.g., such as the CFHT OpenStar Cluster Survey, or the WIYN Open Star ClusterSurvey in the North; Kalirai et al. 2001; Mathieu 2000).Multiband imaging of these co-eval, co-spatial, and iso-metallic systems will provide vital insights into funda-mental stellar evolution. For example, the depth ofLSST will enable construction of luminosity and massfunctions for nearby open clusters down to the hydro-gen burning limit and beyond. Variations in the initialmass function will be studied as a function of environ-ment (e.g., age and metallicity). The wide-field coveragewill also allow us to track how the stellar populations in he Large Synoptic Survey Telescope Figure 28.
The median metallicity map for 2.5 mil-lion main-sequence F-type stars within 10 kpc from the Sun(adapted from Ivezi´c et al. 2008). The metallicity is esti-mated using u − g and g − r colors measured by SDSS. Theposition and size of the mapped region, relative to the rest ofthe Milky Way, is illustrated in the top right corner, wherethe same map is scaled and overlaid on an image of the An-dromeda galaxy. The gradient of the median metallicity isessentially parallel to the Z axis, except in the Monocerosstream region, as marked. LSST will extend this map out to100 kpc, using a sample of over 100 million main-sequence Fstars. each cluster vary as a function of radius, from the core tobeyond the tidal radius. Fainter remnant white dwarfswill be uncovered in both open and globular clusters (thenearest of which are all in the South), thereby providinga crucial link to the properties of the now evolved starsin each system.In summary, the LSST data will revolutionize studiesof the Milky Way and the entire Local Group. We lista few specific Galactic science programs that LSST willenable: • High-resolution studies of the distribution of starsin the outer halo in the six-dimensional spacespanned by position, metallicity and proper mo-tions (e.g., Girard et al. 2006; Bell et al. 2008;Juri´c et al. 2008; Ivezi´c et al. 2008; Bond et al.2010). • The most complete search possible for halostreams, Galaxy satellites and intra-Local Groupstars (e.g. Belokurov et al. 2007a; Walsh et al.2009; Bochanski et al. 2014). k p c SDSS spatial map
Predicted LSST spatial map
20 kpc
Figure 29.
A predicted spatial distribution of stars outto 150 kpc from the center of the Milky Way, from Bul-lock & Johnston (2005). LSST will resolve main sequenceturnoff stars out to 300 kpc, ten times more volume thanshown here, enabling a high-fidelity spatial map over the en-tire observed virial volume. (Note that this is significantlylarger than the 100 kpc probed by metallicity measurementsin Figure 28, which is limited by the depth of the u -bandobservations.) Overlaid on this prediction is the observedSDSS stellar number density map based on main sequencestars with 0 . < r − i < .
15 (Juri´c et al. 2008). This mapextends up to ∼
20 kpc from the Sun, with the white boxshowing a scale of 20 kpc across and the left side aligned withthe Galactic center. The revolutionary Galaxy map providedby SDSS is only complete to ∼
40 kpc, or only ∼
1% of thevirial volume. However, the outermost reaches of the stel-lar halo are predicted to bear the most unique signatures ofour Galaxy’s formation (Johnston et al. 2008; Cooper et al.2010). LSST will be the only survey capable of fully testingsuch predictions. • Deep and highly accurate color-magnitude dia-grams for over half of the known globular clus-ters, including tangential velocities from propermotion measurements (An et al. 2008; Casetti-Dinescu et al. 2007). • Mapping the metallicity, kinematics and spatialprofile of the Saggitarius dwarf tidal stream (e.g.,Ibata et al. 2001; Majewski et al. 2003; Law et al.2005; Belokurov et al. 2014) and the Magellanicstream (Zaritsky et al. 2004). • The measurement of the internal motions of MilkyWay dwarf galaxies via proper motions, therebyconstraining their density profiles and possibly the2
Ivezi´c, Kahn, Tyson, Abel, Acosta, Allsman, Alonso, AlSayyad, Anderson, et al. nature of dark matter (e.g., Walker & Pe˜narrubia2011). • Detailed constraints on the formation and evolu-tion of the populations within the Galactic Bulge,as traced by the spatial distribution, motion, andchemistry of ∼ − of its stars (e.g. Hill et al.2011; Ness et al. 2014). • Studies of the clumpiness of the gravitational po-tential in the Galaxy using fragile wide-angle bi-naries selected with the aid of trigonometric andphotometric parallaxes, and common proper mo-tion (e.g., Yoo et al. 2004; Longhitano & Binggeli2010). • Detailed studies of variable star populations; 2%or better accurate multicolor light curves will beavailable for a sample of at least 50 million variablestars (Sesar et al. 2007), enabling studies of cat-aclysmic variables, eclipsing binary systems, andrare types of variables. • Discovery of rare and faint high proper motion ob-jects: probing the faint end of the stellar massfunction (L´epine 2008; Finch et al. 2010), andsearching for free-floating planet candidates (Lu-cas & Roche 2000; Luhman 2014). • Direct measurement of the faint end of the stellarluminosity function using trigonometric parallaxes(Reid et al. 2002) and a complete census of the so-lar neighborhood to a distance of 100 pc basedon trigonometric parallax measurements for ob-jects as faint as M r = 17 ( ∼ L5 brown dwarfs).For example, LSST will deliver 10% or betterdistances for a sample of about 2,500 stars with18 < M r < • The separation of halo M sub-dwarfs from disk Mdwarfs, using the z − y color which is sensitive totheir rich molecular band structure (West et al.2011; Bochanski et al. 2013). • Studies of white dwarfs using samples of severalmillion objects, including the determination ofthe halo white dwarf luminosity function (SciBookCh. 6). • Measurements of physical properties of stars us-ing large samples of eclipsing binary stars (Stassunet al. 2013). • High-resolution three-dimensional studies of inter-stellar dust using 5-color SEDs of main sequence
Figure 30.
A comparison of an SDSS image (2 × gri composite) showing a typical galaxy at a redshift of ∼ BV R composite image of the same fieldobtained by the MUSYC survey (bottom; Gawiser et al.2006). The MUSYC image is about 4 mag deeper than theSDSS image (and about 1 mag less deep than the anticipatedLSST 10-year coadded data). Note the rich surface bright-ness structure seen in the MUSYC image that is undetectablein the SDSS image. stars (Planck Collaboration et al. 2011; Berry et al.2012; Green et al. 2014). • A census of AGB stars in the Galaxy by searchingfor resolved envelopes and optical identifications ofIR counterparts (e.g., from the WISE survey), andby using long-term variability and color selection(Ivezi´c 2007). • A complete census of faint populations in nearbystar forming regions using color and variability se-lection (e.g. Brice˜no et al. 2005).4.5.
Additional Science Projects
The experience with any large survey (e.g., SDSS,2MASS, VISTA, WISE, GALEX, to name but a few)is that much of their most interesting science is often he Large Synoptic Survey Telescope • Detailed studies of galaxy formation and evolu-tion using their distribution in luminosity-color-morphology space as a function of redshift. For ex-ample, LSST will enable studies of the rest-frameUV emission, similar to those based on GALEXdata for local galaxies, to a redshift of ∼ – the evolution of the galaxy luminosity func-tion with redshift, as a function of morphol-ogy and color; – the evolution of the galaxy color distributionover a wide range of rest-frame wavelengths,and as a function of luminosity and morphol-ogy; – bulge-disk decomposition, as a function ofluminosity and color, over a large redshiftrange; – detailed distribution of satellite galaxies inluminosity-color-morphology space as a func-tion of luminosity, color, and morphology ofthe primary galaxy; – correlations of luminosity, color and mor-phology with local environment from kpc toMpc scales, and as a function of redshift (seeFigs. 30 and 31); – the properties of galaxy groups and clustersas a function of cosmic time. • AGN census to very faint luminosity and a largeredshift limit (Ivezi´c et al. 2014), yielding 20 mil-lion objects from LSST data alone, and the abilityto identify up to ∼
100 million objects once multi-wavelength data are used to aid AGN selection (seeFig. 32). By reaching substantially further downthe AGN luminosity function than has been possi-ble before over a very large solid angle, LSST datawill test evolutionary cosmic downsizing scenar-ios across the full range of cosmic environments,and lead to a much clearer understanding of black-hole growth during the first Gyr. For example,LSST should discover several thousand AGNs at z ∼ − .
5, representing a dramatic increase overpresent samples (Brandt & LSST Active Galax-ies Science Collaboration 2007, see also SciBookCh. 10).
Figure 31.
A comparison of the distribution of galaxiesin luminosity–color–density space measured by SDSS (left)and a model based on the Millennium simulation (right).The linearly-spaced contours outline the distribution of avolume-limited sample of galaxies in the plotted diagram,and the color-coded background shows the median envi-ronmental density (computed using the ten nearest neigh-bors) for galaxies with the corresponding luminosity andcolor. Such multi-variate distributions encode rich informa-tion about formation and evolution of galaxies. Galaxiesdetected by SDSS are representative of the low-redshift Uni-verse (the median redshift is ∼ z ∼
2. Adapted fromCowan & Ivezi´c (2008). • The combination of LSST, Euclid and WFIRSTdata should allow discovery of at least tens ofquasars at z > . • LSST data will provide good constraints on AGNlifetimes, or at least the timescales over whichthey make distinct accretion-state transitions(MacLeod et al. 2016), due to large sample size andsurvey lifetime (e.g. Martini & Schneider 2003). • The first wide field survey of ultra low surfacebrightness galaxies, with photometric redshift in-formation. The currently available samples (e.g.Greco et al. 2018) are highly incomplete, espe-cially in the Southern Hemisphere (see Fig. 7 inBelokurov et al. 2007a). • Search for strong gravitational lenses to a faint sur-face brightness limit (e.g. Bartelmann et al. 1998;Tyson et al. 1998; Belokurov et al. 2007b), whichcan be used to explore the dark matter profiles ofgalaxies (e.g., Treu et al. 2006).4.5.1.
Synergy with other projects
LSST will not operate in isolation and will greatlybenefit from other precursor and coeval data at a vari-ety of wavelengths, depths, and timescales. For exam-ple, in the visual wavelength range, most of the Celes-tial Sphere will be covered to a limit several magnitudes4
Ivezi´c, Kahn, Tyson, Abel, Acosta, Allsman, Alonso, AlSayyad, Anderson, et al.
Figure 32.
The LSST will deliver AGN sky densities of1000–4000 deg − (top panel); The total LSST AGN yield,selected using colors and variability, should be well over 10million objects, especially once multiwavelength data are alsoutilized. The bottom panel shows the expected distributionof these objects in the absolute magnitude vs. redshift plane,color-coded by the probability for an object to be detected asvariable after 1 year of observations. Note that quasars willbe detected to their formal luminosity cutoff ( M < −
23) evenat redshifts of ∼
5. Adapted from Brandt & LSST ActiveGalaxies Science Collaboration (2007). fainter than LSST saturation ( r ∼ ∼ a magni-tude deeper still in the southern sky. Despite the lack ofthe u band data and its relatively shallow imaging, thePan-STARRS surveys represents a valuable complementto LSST in providing Northern sky coverage to a limitfainter than that of SDSS and SkyMapper. LSST andGaia will be highly complementary datasets for study-ing the Milky Way in the multi-dimensional space ofthree-dimensional positions, proper motions and metal-licity (Ivezi´c et al. 2012). The Gaia survey will providecalibration checks at the bright end for proper motionsand trigonometric parallax measurements by LSST, andLSST will extend the Gaia survey by four magnitudes.The upcoming Zwicky Transient Facility (e.g., Laheret al. 2017), with its 600 Megapixel camera and a 47deg large field of view, will generate the largest opti-cal transient stream prior to LSST (at about one tenthof LSST rate) and thus provide an early insight intoastrophysical surprises and technical challenges await-ing LSST. The LSST data stream will invigorate subse-quent investigations by numerous other telescopes thatwill provide additional temporal, spectral and spatialresolution coverage.WFIRST and Euclid will carry out wide-field imagingsurveys in the near-infrared, giving highly complemen-tary photometry to LSST. The resulting galaxy SEDsshould give rise to even better photometric redshifts, aswell as tighter constraints on stellar masses and starformation histories crucial for galaxy evolution stud-ies. The weak lensing analyses from space and fromthe ground will also be highly complementary, and willprovide crucial cross-checks of one another. LSST alsopresents the opportunity to conduct simultaneous obser-vations of WFIRST’s Galactic Bulge survey fields, fromwhich it will be possible to measure the parallax andhence the lens masses for most microlensing events, aswell as providing valuable lightcurve coverage during thelarge data gaps between WFIRST survey ‘seasons’.LSST will also enable multi-wavelength studies of faintoptical sources using gamma-ray, X-ray, IR and radiodata. For example, the SDSS detected only 1/3 of all20cm FIRST sources (Becker et al. 1995) because it wastoo shallow by ∼ Fermi
Gamma-ray Space Telescope (e.g., Atwood et al. 2009). TheLaser Interferometer Gravitational Wave Observatory he Large Synoptic Survey Telescope COMMUNITY INVOLVEMENTLSST has been conceived as a public facility: thedatabase that it will produce, and the associated ob-ject catalogs that are generated from that database,will be made available with no proprietary period tothe U.S. and Chilean scientific communities, as well asto those international partners who contribute to oper-ations funding. As described in §
6, data will also bemade available to the general public for educational andoutreach activities. The LSST data management sys-tem ( § §§ Ivezi´c, Kahn, Tyson, Abel, Acosta, Allsman, Alonso, AlSayyad, Anderson, et al. nical topics of interest to both the science communityand the LSST Project. The SAC minutes and notes areavailable publicly. Current members on this committeeare: T. Anguita (Andr´es Bello, Chile), R. Bean (Cor-nell), W.N. Brandt (Penn State), J. Kalirai (STScI),M. Kasliwal (Caltech), D. Kirkby (UC Irvine), C. Liu(Staten Island), A. Mainzer (JPL), R. Malhotra (UArizona), N. Padilla (U. Cat´olica de Chile), J. Simon(Carnegie), A. Slosar (Brookhaven), M. Strauss (Prince-ton), L. Walkowicz (Adler), and R. Wechsler (Stanford). EDUCATIONAL AND SOCIETAL IMPACTSThe impact and enduring societal significance ofLSST will exceed its direct contributions to advances inphysics and astronomy. LSST is uniquely positioned tohave high impact with the interested public, planetar-iums and science centers, and citizen science projects,as well as middle school through university educationalprograms. LSST will contribute to the national goalsof enhancing science literacy and increasing the globalcompetitiveness of the US science and technology work-force. Engaging the public in LSST activities has beenpart of the project design from the beginning.The mission of LSST’s Education and Public Out-reach (EPO) program is to provide worldwide access to asubset of LSST data through accessible and engaging on-line experiences so anyone can explore the universe andbe part of the discovery process. To do this, LSST EPOwill facilitate a pathway from entry-level exploration ofastronomical imagery and information to more sophisti-cated interaction with LSST data using tools similar towhat professional astronomers use for their work.A dynamic, immersive web portal will enable mem-bers of the public to explore color images of the fullLSST sky, examine objects in more detail, view eventsfrom the nightly alert stream, learn more about LSSTscience topics and discoveries, and investigate scientificquestions that excite them using real LSST data in on-line science notebooks. The portal will also link to nu-merous citizen science projects using LSST data.LSST data can become a key part of classrooms em-phasizing student-centered research in middle school,high school, and undergraduate settings. Using onlinescience notebooks, teachers will be able to bring realLSST telescope data into their classrooms without hav-ing to download, install, and maintain software locally.Educational investigations will be designed to supportkey aspects of the Next-Generation Science Standards(NGSS) in the USA, and goals of the Explora pro-gram through CONICYT in Chile. Educators will besupported through professional development that offerstraining on the online notebook technology and also rel- evant science content. Science notebooks will also ac-commodate access to LSST data for lifelong learners andanyone that visits the portal.Anyone around the world will be able to participatein a variety of citizen science projects that use LSSTdata. The EPO Team will work with the Zooniverse todevelop the
Project Builder to include tools specificallydesigned to utilize LSST data, allowing LSST principalinvestigators to create any number of projects to helpthem accomplish their science goals. EPO anticipatesthat the number of citizen science projects in the as-tronomy field will increase dramatically when LSST isoperational, giving a whole new generation of citizen sci-entists the opportunity to deepen their engagement withastronomy using authentic data from LSST.LSST EPO will produce and maintain a digital libraryof multimedia assets including images, video clips, and3D models. Multimedia assets will be aligned to stan-dards such as IMERSA Dome Master and AstronomyVisualization Metadata, when applicable, allowing fullflexibility for adoption by content creators at planetari-ums and science centers. We will also follow the Inter-national Planetarium Society’s Data2Dome standard, tomaximize the number of platforms that can use our as-sets.The LSST EPO program will rely on a cloud-basedEPO Data Center (EDC) to handle the unique needs ofthe EPO audiences. These needs include, for example,a fast and smooth browsing experience on mobile de-vices, and the need to handle inevitable spikes and lullsin visitor traffic and data transfers. As such, the EDCwill follow best practices popularized by cloud comput-ing, leveraging on-demand computing and auto-scalablearchitecture. Remaining agile and relevant during thefull lifetime of Operations by adjusting to technologytrends and education priorities is an important part ofthe LSST EPO design process.LSST EPO is committed to engaging with diverse au-diences and is undertaking a multi-faceted approach toreaching diverse individuals. LSST EPO is planningto partner with at least five organizations serving 1)women/girls, 2) individuals from traditionally underrep-resented groups in STEM, and 3) individuals in low so-cioeconomic communities. Representatives from theseorganizations will be key stakeholders of the EPO pro-gram, helping to shape deliverables and a culturally re-sponsive program evaluation. Furthermore, these rela-tionships will allow for co-creation of EPO deliverablesto help ensure materials are accessible to, of interest to,and relevant to diverse populations.LSST EPO is breaking new ground in bringing astro-nomical data to the public in a timely, engaging, and he Large Synoptic Survey Telescope SUMMARY AND CONCLUSIONSUntil recently, most astronomical investigations havefocused on small samples of cosmic sources or individ-ual objects. Over the past few decades, however, ad-vances in technology have made it possible to move be-yond the traditional observational paradigm and to un-dertake large-scale sky surveys, such as SDSS, 2MASS,GALEX, Gaia, and others. This observational progress,based on a synergy of advances in telescope construc-tion, detectors, and above all, information technology,has had a dramatic impact on nearly all fields of astron-omy, many areas of fundamental physics, and society ingeneral. The LSST builds on the experience of these sur-veys and addresses the broad goals stated in several na-tionally endorsed reports by the U.S. National Academyof Sciences. The 2010 report “New Worlds, New Hori-zons in Astronomy and Astrophysics” by the Commit-tee for a Decadal Survey of Astronomy and Astrophysicsranked LSST as its top priority for large ground-basedprograms. The LSST will be unique: the combinationof large aperture and large field of view, coupled withthe needed computation power and database technology,will enable simultaneously fast and wide and deep imag-ing of the sky, addressing in one sky survey the broadscientific community’s needs in both the time domainand deep universe.The realization of the LSST involves extraordinary en-gineering and technological challenges: the fabricationof large, high-precision optics; construction of a huge,highly-integrated array of sensitive, wide-band imagingsensors; and the operation of a data management facilityhandling tens of terabytes of data each day. The design,development and construction effort has been underwaysince 2006 and will continue through the onset of fullsurvey operations. This work involves hundreds of per-sonnel at institutions all over the US, Chile, and the restof the world.In December 2013, LSST passed the NSF Final De-sign Review for construction, and in May 2014 the Na-tional Science Board approved the project. The pri-mary/tertiary mirror was cast in 2008, and the polishedmirror was completed in 2015. In 2014 LSST transi-tioned from the design and development phase to con-struction, and the Associated Universities for Research in Astronomy (AURA) has formal responsibility for theLSST project since 2011. At this writing, the project isnear the peak of the construction effort, and is preparingfor the transition to commissioning and operations.The construction cost of LSST is being borne by theUS National Science Foundation, the Department of En-ergy, generous contributions from several private foun-dations and institutions, and the member institutions ofthe LSST Corporation. The LSST budget includes a sig-nificant Education and Public Outreach program ( § Ivezi´c, Kahn, Tyson, Abel, Acosta, Allsman, Alonso, AlSayyad, Anderson, et al.
This material is based upon work supported in partby the National Science Foundation through Coopera-tive Agreement 1258333 managed by the Association ofUniversities for Research in Astronomy (AURA), andthe Department of Energy under Contract No. DE-AC02-76SF00515 with the SLAC National Accelerator Laboratory. Additional LSST funding comes from pri-vate donations, grants to universities, and in-kind sup-port from LSSTC Institutional Members.
Facility:
LSSTAPPENDIX A. VERSION HISTORYVersion 1.0 (May 15, 2008): the first posting.Version 2.0 (June 7, 2011): acknowledged the Decadal Survey 2010 report; updated construction schedule; updatedexpected performance in Table 2; added sections on LSST simulations and data mining; updated several figures;updated references; expanded author list.Version 3.0 (August 26, 2014): acknowledged the start of federal construction; updated system description andscience examples, updated several figures; refreshed references; expanded author list.Version 4.0 (May 15, 2018): updated system description and science examples, updated expected performance inTable 2; updated several figures; refreshed references; expanded author list.REFERENCES
Abazajian, K. N., Adelman-McCarthy, J. K., Ag¨ueros,M. A., et al. 2009, ApJS, 182, 543,doi: 10.1088/0067-0049/182/2/543Abazajian, K. N., Adshead, P., Ahmed, Z., et al. 2016,ArXiv e-prints. https://arxiv.org/abs/1610.02743Abbott, B. P., Abbott, R., Abbott, T. D., et al. 2017a,ApJL, 848, L12, doi: 10.3847/2041-8213/aa91c9—. 2017b, Nature, 551, 85, doi: 10.1038/nature24471A’Hearn, M. F. 2004, in Comets II, ed. M. C. Festou, H. U.Keller, & H. A. Weaver (University of Arizona Press),17–22A’Hearn, M. F., Millis, R. C., Schleicher, D. O., Osip, D. J.,& Birch, P. V. 1995, Icarus, 118, 223,doi: 10.1006/icar.1995.1190Aihara, H., Arimoto, N., Armstrong, R., et al. 2018, PASJ,70, S4, doi: 10.1093/pasj/psx066Albrecht, A., Bernstein, G., Cahn, R., et al. 2006, ArXivAstrophysics e-prints.https://arxiv.org/abs/astro-ph/0609591Alcock, C., Allsman, R. A., Alves, D. R., et al. 2000, ApJ,542, 281, doi: 10.1086/309512—. 2001, ApJL, 550, L169, doi: 10.1086/319636Amˆores, E. B., & L´epine, J. R. D. 2005, AJ, 130, 659,doi: 10.1086/430957An, D., Johnson, J. A., Clem, J. L., et al. 2008, ApJS, 179,326, doi: 10.1086/592090An, D., Beers, T. C., Johnson, J. A., et al. 2013, ApJ, 763,65, doi: 10.1088/0004-637X/763/1/65 Anderson, G. P., Berk, A., Acharya, P. K., et al. 1999, inProc. SPIE, Vol. 3866, Optics in AtmosphericPropagation and Adaptive Systems III, ed. A. Kohnle &J. D. Gonglewski, 2–10Anderson, S. F., Haggard, D., Homer, L., et al. 2005, AJ,130, 2230, doi: 10.1086/491587Angel, R., Lesser, M., Sarlot, R., & Dunham, E. 2000, inASP Conf. Ser., Vol. 195, Imaging the Universe in ThreeDimensions, ed. W. van Breugel & J. Bland-Hawthorn, 81Angeli, G. Z., Xin, B., Claver, C., et al. 2016, inProc. SPIE, Vol. 9911, Modeling, Systems Engineering,and Project Management for Astronomy VI, 991118Antilogus, P., Bailly, P., Barrillon, P., et al. 2017, Journalof Instrumentation, 12, C03017Antoniou, I., & Perivolaropoulos, L. 2010, JCAP, 12, 012,doi: 10.1088/1475-7516/2010/12/012Araujo-Hauck, C., Sebag, J., Liang, M., et al. 2016, inProc. SPIE, Vol. 9906, Ground-based and AirborneTelescopes VI, 99060LArcavi, I., Howell, D. A., Kasen, D., et al. 2017, Nature,551, 210, doi: 10.1038/nature24030Arnett, W. D., & Meakin, C. 2011, ApJ, 741, 33,doi: 10.1088/0004-637X/741/1/33Astropy Collaboration, Robitaille, T. P., Tollerud, E. J.,et al. 2013, A&A, 558, A33,doi: 10.1051/0004-6361/201322068Atwood, W. B., Abdo, A. A., Ackermann, M., et al. 2009,ApJ, 697, 1071, doi: 10.1088/0004-637X/697/2/1071 he Large Synoptic Survey Telescope Bacci, P., Maestripieri, M., Tesi, L., et al. 2017, MinorPlanet Electronic Circulars, 2017-U181Barnard, M., Abrahamse, A., Albrecht, A., Bozek, B., &Yashar, M. 2008, PhRvD, 78, 043528,doi: 10.1103/PhysRevD.78.043528Bartelmann, M., Huss, A., Colberg, J. M., Jenkins, A., &Pearce, F. R. 1998, A&A, 330, 1Batista, V., Beaulieu, J.-P., Gould, A., et al. 2014, ApJ,780, 54, doi: 10.1088/0004-637X/780/1/54Batygin, K., & Brown, M. E. 2016, AJ, 151, 22,doi: 10.3847/0004-6256/151/2/22Bayliss, D. D. R., & Sackett, P. D. 2011, ApJ, 743, 103,doi: 10.1088/0004-637X/743/2/103Beaulieu, J.-P., Bennett, D. P., Fouqu´e, P., et al. 2006,Nature, 439, 437, doi: 10.1038/nature04441Becker, A. C., Wittman, D. M., Boeshaar, P. C., et al.2004, ApJ, 611, 418, doi: 10.1086/421994Becker, R. H., White, R. L., & Helfand, D. J. 1995, ApJ,450, 559, doi: 10.1086/176166Becla, J., & Wang, D. L. 2014, in Exascale RadioAstronomy, Vol. 2Becla, J., Wang, D., Monkewitz, S., et al. 2017, LSST DataManagement Database Design, LSST DocumentLDM-135, LSST. https://ls.st/LDM-135Bell, E. F., Zucker, D. B., Belokurov, V., et al. 2008, ApJ,680, 295, doi: 10.1086/588032Bellm, E. 2014, in The Third Hot-wiring the TransientUniverse Workshop, ed. P. R. Wozniak, M. J. Graham,A. A. Mahabal, & R. Seaman, 27–33Belokurov, V., & Koposov, S. E. 2016, MNRAS, 456, 602,doi: 10.1093/mnras/stv2688Belokurov, V., Zucker, D. B., Evans, N. W., et al. 2007a,ApJ, 654, 897, doi: 10.1086/509718Belokurov, V., Evans, N. W., Moiseev, A., et al. 2007b,ApJL, 671, L9, doi: 10.1086/524948Belokurov, V., Koposov, S. E., Evans, N. W., et al. 2014,MNRAS, 437, 116, doi: 10.1093/mnras/stt1862Benson, B. A., Ade, P. A. R., Ahmed, Z., et al. 2014, inProc. SPIE, Vol. 9153, Millimeter, Submillimeter, andFar-Infrared Detectors and Instrumentation forAstronomy VII, 91531PBergeron, P., Wesemael, F., & Beauchamp, A. 1995, PASP,107, 1047, doi: 10.1086/133661Bernstein, G. M., Trilling, D. E., Allen, R. L., et al. 2004,AJ, 128, 1364, doi: 10.1086/422919Bernstein, G. M., Armstrong, R., Plazas, A. A., et al. 2017,PASP, 129, 074503, doi: 10.1088/1538-3873/aa6c55Berry, M., Ivezi´c, ˇZ., Sesar, B., et al. 2012, ApJ, 757, 166,doi: 10.1088/0004-637X/757/2/166 Bhattacharya, S., Kosowsky, A., Newman, J. A., &Zentner, A. R. 2011, PhRvD, 83, 043004,doi: 10.1103/PhysRevD.83.043004Bianco, F. B., Modjaz, M., Hicken, M., et al. 2014, ApJS,213, 19, doi: 10.1088/0067-0049/213/2/19Blake, C., & Glazebrook, K. 2003, ApJ, 594, 665,doi: 10.1086/376983Blandford, R. D., & LSST Strong Lensing ScienceCollaboration. 2007, in Bulletin of the AmericanAstronomical Society, Vol. 39, American AstronomicalSociety Meeting Abstracts, 979Bloom, J. S., Starr, D. L., Butler, N. R., et al. 2008,Astronomische Nachrichten, 329, 284,doi: 10.1002/asna.200710957Bloom, J. S., Giannios, D., Metzger, B. D., et al. 2011,Science, 333, 203, doi: 10.1126/science.1207150Bochanski, J. J., Savcheva, A., West, A. A., & Hawley,S. L. 2013, AJ, 145, 40, doi: 10.1088/0004-6256/145/2/40Bochanski, J. J., West, A. A., Hawley, S. L., & Covey,K. R. 2007, AJ, 133, 531, doi: 10.1086/510240Bochanski, J. J., Willman, B., West, A. A., Strader, J., &Chomiuk, L. 2014, AJ, 147, 76,doi: 10.1088/0004-6256/147/4/76Bockel´ee-Morvan, D., & Biver, N. 2017, PhilosophicalTransactions of the Royal Society of London Series A,375, 20160252, doi: 10.1098/rsta.2016.0252Bond, N. A., Ivezi´c, ˇZ., Sesar, B., et al. 2010, ApJ, 716, 1,doi: 10.1088/0004-637X/716/1/1Bongiorno, A., Zamorani, G., Gavignaud, I., et al. 2007,A&A, 472, 443, doi: 10.1051/0004-6361:20077611Bonvin, V., Courbin, F., Suyu, S. H., et al. 2017, MNRAS,465, 4914, doi: 10.1093/mnras/stw3006Bosch, J., Armstrong, R., Bickerton, S., et al. 2018, PASJ,70, S5, doi: 10.1093/pasj/psx080Bottke, W. F., Vokrouhlick´y, D., Broz, M., Nesvorn´y, D., &Morbidelli, A. 2001, Science, 294, 1693,doi: 10.1126/science.1066760Brandt, N., & LSST Active Galaxies Science Collaboration.2007, in Bulletin of the American Astronomical Society,Vol. 39, American Astronomical Society MeetingAbstracts, 980Brandt, W. N., & Hasinger, G. 2005, ARA&A, 43, 827,doi: 10.1146/annurev.astro.43.051804.102213Brandt, W. N., & Vito, F. 2017, AstronomischeNachrichten, 338, 241, doi: 10.1002/asna.201713337Brice˜no, C., Calvet, N., Hern´andez, J., et al. 2005, AJ, 129,907, doi: 10.1086/426911Brown, M. E. 2017, AJ, 154, 65,doi: 10.3847/1538-3881/aa79f4 Ivezi´c, Kahn, Tyson, Abel, Acosta, Allsman, Alonso, AlSayyad, Anderson, et al.
Brown, M. E., & Batygin, K. 2016, ApJL, 824, L23,doi: 10.3847/2041-8205/824/2/L23Brown, M. E., Trujillo, C., & Rabinowitz, D. 2004, ApJ,617, 645, doi: 10.1086/422095Bruzual, G., & Charlot, S. 2003, MNRAS, 344, 1000,doi: 10.1046/j.1365-8711.2003.06897.xBudav´ari, T., Connolly, A. J., Szalay, A. S., et al. 2003,ApJ, 595, 59, doi: 10.1086/377168Bullock, J. S., & Johnston, K. V. 2005, ApJ, 635, 931,doi: 10.1086/497422Burke, D. L., Axelrod, T., Blondin, S., et al. 2010, ApJ,720, 811, doi: 10.1088/0004-637X/720/1/811Burke, D. L., Rykoff, E. S., Allam, S., et al. 2018, AJ, 155,doi: 10.3847/1538-3881/aa9f22Burrows, A., Sudarsky, D., & Hubeny, I. 2006, ApJ, 640,1063, doi: 10.1086/500293Cai, R.-G., Ma, Y.-Z., Tang, B., & Tuo, Z.-L. 2013,PhRvD, 87, 123522, doi: 10.1103/PhysRevD.87.123522Campanelli, L., Cea, P., Fogli, G. L., & Marrone, A. 2011,PhRvD, 83, 103503, doi: 10.1103/PhysRevD.83.103503Carollo, D., Beers, T. C., Lee, Y. S., et al. 2007, Nature,450, 1020, doi: 10.1038/nature06460Casetti-Dinescu, D. I., Girard, T. M., Herrera, D., et al.2007, AJ, 134, 195, doi: 10.1086/518507Chambers, K. C., Magnier, E. A., Metcalfe, N., et al. 2016,ArXiv e-prints. https://arxiv.org/abs/1612.05560Chang, C., Jarvis, M., Jain, B., et al. 2013, MNRAS, 434,2121, doi: 10.1093/mnras/stt1156—. 2015, MNRAS, 447, 1746, doi: 10.1093/mnras/stu2553Chapline, G. F., & Frampton, P. H. 2016, JCAP, 11, 042,doi: 10.1088/1475-7516/2016/11/042Chelouche, D., & Daniel, E. 2012, ApJ, 747, 62,doi: 10.1088/0004-637X/747/1/62Choi, J.-Y., Han, C., Udalski, A., et al. 2013, ApJ, 768,129, doi: 10.1088/0004-637X/768/2/129Chou, M.-Y., Majewski, S. R., Cunha, K., et al. 2007, ApJ,670, 346, doi: 10.1086/522483Coe, D., & Moustakas, L. A. 2009, ApJ, 706, 45,doi: 10.1088/0004-637X/706/1/45Colin, J., Mohayaee, R., Sarkar, S., & Shafieloo, A. 2011,MNRAS, 414, 264, doi: 10.1111/j.1365-2966.2011.18402.xConnolly, A. J., Angeli, G. Z., Chandrasekharan, S., et al.2014, in Proc. SPIE, Vol. 9150, Modeling, SystemsEngineering, and Project Management for Astronomy VI,915014Cook, N. V., Ragozzine, D., Granvik, M., & Stephens, D. C.2016, ApJ, 825, 51, doi: 10.3847/0004-637X/825/1/51Cooper, A. P., Cole, S., Frenk, C. S., et al. 2010, MNRAS,406, 744, doi: 10.1111/j.1365-2966.2010.16740.x Cooray, A., Hu, W., Huterer, D., & Joffre, M. 2001, ApJL,557, L7, doi: 10.1086/323323Cooray, A. R. 1999, A&A, 348, 31Coughlin, E. R., Armitage, P. J., Nixon, C., & Begelman,M. C. 2017a, MNRAS, 465, 3840,doi: 10.1093/mnras/stw2913Coughlin, M., Dietrich, T., Kawaguchi, K., et al. 2017b,ApJ, 849, 12, doi: 10.3847/1538-4357/aa9114Cowan, N. B., & Ivezi´c, ˇZ. 2008, ApJL, 674, L13,doi: 10.1086/528986Cowperthwaite, P. S., & Berger, E. 2015, ApJ, 814, 25,doi: 10.1088/0004-637X/814/1/25Cuadra, J., Armitage, P. J., Alexander, R. D., & Begelman,M. C. 2009, MNRAS, 393, 1423,doi: 10.1111/j.1365-2966.2008.14147.xCuillandre, J.-C. J., Withington, K., Hudelot, P., et al.2012, in Proc. SPIE, Vol. 8448, Observatory Operations:Strategies, Processes, and Systems IV, 84480MCushing, M. C., Rayner, J. T., & Vacca, W. D. 2005, ApJ,623, 1115, doi: 10.1086/428040Dai, D.-C., Kinney, W. H., & Stojkovic, D. 2011, JCAP, 4,015, doi: 10.1088/1475-7516/2011/04/015Dalal, N., Dor´e, O., Huterer, D., & Shirokov, A. 2008,PhRvD, 77, 123514, doi: 10.1103/PhysRevD.77.123514Dalal, N., & Kochanek, C. S. 2002, ApJ, 572, 25,doi: 10.1086/340303Dandy, C. L., Fitzsimmons, A., & Collander-Brown, S. J.2003, Icarus, 163, 363,doi: 10.1016/S0019-1035(03)00087-3Dark Energy Survey Collaboration, Abbott, T., Abdalla,F. B., et al. 2016, MNRAS, 460, 1270,doi: 10.1093/mnras/stw641Davis, C., Gatti, M., Vielzeuf, P., et al. 2017, ArXive-prints. https://arxiv.org/abs/1710.02517Dawson, W. A., Wittman, D., Jee, M. J., et al. 2012, ApJL,747, L42, doi: 10.1088/2041-8205/747/2/L42De Bernardis, F., Stevens, J. R., Hasselfield, M., et al.2016, in Proc. SPIE, Vol. 9910, Observatory Operations:Strategies, Processes, and Systems VI, 991014de Bruijne, J. H. J. 2012, Ap&SS, 341, 31,doi: 10.1007/s10509-012-1019-4de Jong, J. T. A., Kuijken, K. H., & H´eraudeau, P. 2008,A&A, 478, 755, doi: 10.1051/0004-6361:20078378De Lucia, G., Springel, V., White, S. D. M., Croton, D., &Kauffmann, G. 2006, MNRAS, 366, 499,doi: 10.1111/j.1365-2966.2005.09879.xde Vries, W. H., Becker, R. H., & White, R. L. 2003, AJ,126, 1217, doi: 10.1086/377486 he Large Synoptic Survey Telescope Delgado, F., Saha, A., Chandrasekharan, S., et al. 2014, inProc. SPIE, Vol. 9150, Modeling, Systems Engineering,and Project Management for Astronomy VI, 915015della Valle, M., & Livio, M. 1995, ApJ, 452, 704,doi: 10.1086/176342DeMeo, F. E., Binzel, R. P., Slivan, S. M., & Bus, S. J.2009, Icarus, 202, 160, doi: 10.1016/j.icarus.2009.02.005Denneau, L., Jedicke, R., Grav, T., et al. 2013, PASP, 125,357, doi: 10.1086/670337DES Collaboration, Abbott, T. M. C., Abdalla, F. B., et al.2017, ArXiv e-prints. https://arxiv.org/abs/1708.01530Djorgovski, S. G., Mahabal, A., Drake, A., Graham, M., &Donalek, C. 2013, in Planets, Stars and StellarSystems. Volume 2: Astronomical Techniques, Softwareand Data, ed. T. D. Oswalt & H. E. Bond (Springer), 223Djorgovski, S. G., Mahabal, A. A., Drake, A. J., et al. 2012,in IAU Symposium, Vol. 285, New Horizons in TimeDomain Astronomy, ed. E. Griffin, R. Hanisch, &R. Seaman, 141–146Dodelson, S., & Vallinotto, A. 2006, PhRvD, 74, 063515,doi: 10.1103/PhysRevD.74.063515Dolney, D., Jain, B., & Takada, M. 2006, MNRAS, 366,884, doi: 10.1111/j.1365-2966.2005.09606.xDoressoundiram, A., Peixinho, N., Moullet, A., et al. 2007,AJ, 134, 2186, doi: 10.1086/522783Drake, A. J., Djorgovski, S. G., Mahabal, A., et al. 2009,ApJ, 696, 870, doi: 10.1088/0004-637X/696/1/870Drake, A. J., Beshore, E., Catelan, M., et al. 2010, ArXive-prints. https://arxiv.org/abs/1009.3048Drlica-Wagner, A., Albert, A., Bechtol, K., et al. 2015,ApJL, 809, L4, doi: 10.1088/2041-8205/809/1/L4Drout, M. R., Chornock, R., Soderberg, A. M., et al. 2014,ApJ, 794, 23, doi: 10.1088/0004-637X/794/1/23Duncan, M. J., Levison, H. F., & Budd, S. M. 1995, AJ,110, 3073, doi: 10.1086/117748Durech, J., Kaasalainen, M., Warner, B. D., et al. 2009,A&A, 493, 291, doi: 10.1051/0004-6361:200810393Eisenstein, D. J., Hu, W., & Tegmark, M. 1998, ApJL, 504,L57, doi: 10.1086/311582Elliot, J. L., Kern, S. D., Clancy, K. B., et al. 2005, AJ,129, 1117, doi: 10.1086/427395Engelhardt, T., Jedicke, R., Vereˇs, P., et al. 2017, AJ, 153,133, doi: 10.3847/1538-3881/aa5c8aEvans, C. R., & Kochanek, C. S. 1989, ApJL, 346, L13,doi: 10.1086/185567Feindt, U., Kerschhaggl, M., Kowalski, M., et al. 2013,A&A, 560, A90, doi: 10.1051/0004-6361/201321880Finch, C. T., Zacharias, N., & Henry, T. J. 2010, AJ, 140,844, doi: 10.1088/0004-6256/140/3/844 Fitzpatrick, M. J., Graham, M. J., Mighell, K. J., et al.2016, in Proc. SPIE, Vol. 9913, Software andCyberinfrastructure for Astronomy IV, 99130LFoley, R. J. 2012, ApJ, 748, 127,doi: 10.1088/0004-637X/748/2/127Fukugita, M., Ichikawa, T., Gunn, J. E., et al. 1996, AJ,111, 1748, doi: 10.1086/117915Gaia Collaboration, Brown, A. G. A., Vallenari, A., et al.2016, A&A, 595, A2, doi: 10.1051/0004-6361/201629512Gal-Yam, A., Ofek, E. O., Poznanski, D., et al. 2006, ApJ,639, 331, doi: 10.1086/499157Gardner, J. P., Mather, J. C., Clampin, M., et al. 2006,SSRv, 123, 485, doi: 10.1007/s11214-006-8315-7Garnavich, P. M., Smith, R. C., Miknaitis, G., et al. 2004,in Bulletin of the American Astronomical Society,Vol. 36, American Astronomical Society MeetingAbstracts, 1530Gawiser, E., van Dokkum, P. G., Herrera, D., et al. 2006,ApJS, 162, 1, doi: 10.1086/497644Gezari, S. 2012, in European Physical Journal Web ofConferences, Vol. 39, European Physical Journal Web ofConferences, 03001Gezari, S., Basa, S., Martin, D. C., et al. 2008, ApJ, 676,944, doi: 10.1086/529008Gezari, S., Chornock, R., Rest, A., et al. 2012, Nature, 485,217, doi: 10.1038/nature10990Giannios, D., & Metzger, B. D. 2011, MNRAS, 416, 2102,doi: 10.1111/j.1365-2966.2011.19188.xGirard, T. M., Korchagin, V. I., Casetti-Dinescu, D. I.,et al. 2006, AJ, 132, 1768, doi: 10.1086/507331Gladman, B., Kavelaars, J. J., Petit, J.-M., et al. 2001, AJ,122, 1051, doi: 10.1086/322080Goldstein, D. A., & Nugent, P. E. 2017, ApJL, 834, L5,doi: 10.3847/2041-8213/834/1/L5Goldstein, D. A., Nugent, P. E., Kasen, D. N., & Collett,T. E. 2017, ArXiv e-prints.https://arxiv.org/abs/1708.00003Gonz´alez, J. E., Lacey, C. G., Baugh, C. M., Frenk, C. S.,& Benson, A. J. 2009, MNRAS, 397, 1254,doi: 10.1111/j.1365-2966.2009.15057.xGordon, C., Land, K., & Slosar, A. 2007, Physical ReviewLetters, 99, 081301, doi: 10.1103/PhysRevLett.99.081301Gould, A. 1992, ApJ, 392, 442, doi: 10.1086/171443Graham, M. J., Djorgovski, S. G., Drake, A. J., et al. 2014,MNRAS, 439, 703, doi: 10.1093/mnras/stt2499Graham, M. L., Connolly, A. J., Ivezi´c, ˇZ., et al. 2018, AJ,155, 1, doi: 10.3847/1538-3881/aa99d4Granvik, M., Virtanen, J., Oszkiewicz, D., & Muinonen, K.2009, Meteoritics and Planetary Science, 44, 1853,doi: 10.1111/j.1945-5100.2009.tb01994.x Ivezi´c, Kahn, Tyson, Abel, Acosta, Allsman, Alonso, AlSayyad, Anderson, et al.
Graur, O., Bianco, F. B., Huang, S., et al. 2017, ApJ, 837,120, doi: 10.3847/1538-4357/aa5eb8Grav, T., Jedicke, R., Denneau, L., et al. 2007, in Bulletinof the American Astronomical Society, Vol. 39, AmericanAstronomical Society Meeting Abstracts, 807Grav, T., Mainzer, A. K., & Spahr, T. 2016, AJ, 151, 172,doi: 10.3847/0004-6256/151/6/172Greco, J. P., Greene, J. E., Strauss, M. A., et al. 2018, ApJ,857, doi: 10.3847/1538-4357/aab842Green, G. M., Schlafly, E. F., Finkbeiner, D. P., et al. 2014,ApJ, 783, 114, doi: 10.1088/0004-637X/783/2/114Grier, C. J., Trump, J. R., Shen, Y., et al. 2017, ApJ, 851,21, doi: 10.3847/1538-4357/aa98dcGunn, J. E., Carr, M., Rockosi, C., et al. 1998, AJ, 116,3040, doi: 10.1086/300645Gunn, J. E., Siegmund, W. A., Mannery, E. J., et al. 2006,AJ, 131, 2332, doi: 10.1086/500975Habibi, F., Moniez, M., Ansari, R., & Rahvar, S. 2011,A&A, 525, A108, doi: 10.1051/0004-6361/201015260Han, C. 2008, ApJ, 681, 806, doi: 10.1086/588083Hannestad, S., Tu, H., & Wong, Y. Y. 2006, JCAP, 6, 025,doi: 10.1088/1475-7516/2006/06/025Hargis, J. R., Willman, B., & Peter, A. H. G. 2014, ApJL,795, L13, doi: 10.1088/2041-8205/795/1/L13Harrison, F. A., Craig, W. W., Christensen, F. E., et al.2013, ApJ, 770, 103, doi: 10.1088/0004-637X/770/2/103Hartman, J. D., Gaudi, B. S., Holman, M. J., et al. 2009,ApJ, 695, 336, doi: 10.1088/0004-637X/695/1/336Helmi, A., Ivezi´c, ˇZ., Prada, F., et al. 2003, ApJ, 586, 195,doi: 10.1086/367536Hicken, M., Wood-Vasey, W. M., Blondin, S., et al. 2009,ApJ, 700, 1097, doi: 10.1088/0004-637X/700/2/1097Hill, V., Lecureur, A., G´omez, A., et al. 2011, A&A, 534,A80, doi: 10.1051/0004-6361/200913757H¨oflich, P., Wheeler, J. C., & Thielemann, F. K. 1998, ApJ,495, 617, doi: 10.1086/305327Hovatta, T., Pavlidou, V., King, O. G., et al. 2014,MNRAS, 439, 690, doi: 10.1093/mnras/stt2494Howell, D. A., Sullivan, M., Conley, A., & Carlberg, R.2007, ApJL, 667, L37, doi: 10.1086/522030Howlett, C., Robotham, A. S. G., Lagos, C. D. P., & Kim,A. G. 2017, ApJ, 847, 128,doi: 10.3847/1538-4357/aa88c8Hsieh, H. H., & Jewitt, D. 2006, Science, 312, 561,doi: 10.1126/science.1125150Hu, W., & Haiman, Z. 2003, PhRvD, 68, 063004,doi: 10.1103/PhysRevD.68.063004Hu, W., & Jain, B. 2004, PhRvD, 70, 043009,doi: 10.1103/PhysRevD.70.043009 Hu, W., & Kravtsov, A. V. 2003, ApJ, 584, 702,doi: 10.1086/345846Hu, W., & Tegmark, M. 1999, ApJL, 514, L65,doi: 10.1086/311947Huterer, D., Shafer, D. L., & Schmidt, F. 2015, JCAP, 12,033, doi: 10.1088/1475-7516/2015/12/033Huterer, D., Takada, M., Bernstein, G., & Jain, B. 2006,MNRAS, 366, 101, doi: 10.1111/j.1365-2966.2005.09782.xIbata, R., Irwin, M., Lewis, G. F., & Stolte, A. 2001, ApJL,547, L133, doi: 10.1086/318894Ishak, M., Upadhye, A., & Spergel, D. N. 2006, PhRvD, 74,043513, doi: 10.1103/PhysRevD.74.043513Ivezi´c, ˇZ. 2007, in ASP Conf. Ser., Vol. 378, Why GalaxiesCare About AGB Stars: Their Importance as Actors andProbes, ed. F. Kerschbaum, C. Charbonnel, & R. F.Wing, 485Ivezi´c, ˇZ., Beers, T. C., & Juri´c, M. 2012, ARA&A, 50, 251,doi: 10.1146/annurev-astro-081811-125504Ivezi´c, ˇZ., Brandt, W. N., Fan, X., et al. 2014, in IAUSymposium, Vol. 304, Multiwavelength AGN Surveys andStudies, ed. A. M. Mickaelian & D. B. Sanders, 11–17Ivezi´c, ˇZ., & The LSST Science Collaboration. 2011, LSSTScience Requirements Document, LSST DocumentLPM-17, LSST. https://ls.st/LPM-17Ivezi´c, ˇZ., Tabachnik, S., Rafikov, R., et al. 2001, AJ, 122,2749, doi: 10.1086/323452Ivezi´c, ˇZ., Lupton, R., Schlegel, D., et al. 2004, in ASPConf. Ser., Vol. 327, Satellites and Tidal Streams, ed.F. Prada, D. Martinez Delgado, & T. J. Mahoney, 104Ivezi´c, ˇZ., Tyson, J. A., Juri´c, M., et al. 2007a, in IAUSymposium, Vol. 236, Near Earth Objects, our CelestialNeighbors: Opportunity and Risk, ed. G. B. Valsecchi,D. Vokrouhlick´y, & A. Milani, 353–362Ivezi´c, ˇZ., Smith, J. A., Miknaitis, G., et al. 2007b, AJ, 134,973, doi: 10.1086/519976Ivezi´c, ˇZ., Sesar, B., Juri´c, M., et al. 2008, ApJ, 684, 287,doi: 10.1086/589678Jain, B., Vikram, V., & Sakstein, J. 2013, ApJ, 779, 39,doi: 10.1088/0004-637X/779/1/39Jain, B., & Zhang, P. 2008, PhRvD, 78, 063503,doi: 10.1103/PhysRevD.78.063503Jain, B., Spergel, D., Bean, R., et al. 2015, ArXiv e-prints.https://arxiv.org/abs/1501.07897Javanmardi, B., Porciani, C., Kroupa, P., &Pflamm-Altenburg, J. 2015, ApJ, 810, 47,doi: 10.1088/0004-637X/810/1/47Jedicke, R., Denneau, L., Grav, T., et al. 2005, in Bulletinof the American Astronomical Society, Vol. 37, AmericanAstronomical Society Meeting Abstracts, 1363 he Large Synoptic Survey Telescope Jedicke, R., Nesvorn´y, D., Whiteley, R., Ivezi´c, ˇZ., & Juri´c,M. 2004, Nature, 429, 275, doi: 10.1038/nature02578Jee, M. J., & Tyson, J. A. 2011, PASP, 123, 596,doi: 10.1086/660137Jenness, T., Bosch, J., Owen, R., et al. 2016, in Proc. SPIE,Vol. 9913, Software and Cyberinfrastructure forAstronomy IV, 99130GJenness, T., Economou, F., Findeisen, K., et al. 2018, inProc. SPIE, Vol. 10707, Software andCyberinfrastructure for Astronomy V, in press.https://ls.st/Document-28452Jewitt, D. 2012, AJ, 143, 66,doi: 10.1088/0004-6256/143/3/66Jewitt, D., Stuart, J. S., & Li, J. 2011, AJ, 142, 28,doi: 10.1088/0004-6256/142/1/28Jewitt, D. C., Trujillo, C. A., & Luu, J. X. 2000, AJ, 120,1140, doi: 10.1086/301453Jiang, Y.-F., Green, P. J., Greene, J. E., et al. 2017, ApJ,836, 186, doi: 10.3847/1538-4357/aa5b91Johnston, K. V., Bullock, J. S., Sharma, S., et al. 2008,ApJ, 689, 936, doi: 10.1086/592228Jones, R. L., Gladman, B., Petit, J.-M., et al. 2006, Icarus,185, 508, doi: 10.1016/j.icarus.2006.07.024Jones, R. L., Chesley, S. R., Connolly, A. J., et al. 2007, inBulletin of the American Astronomical Society, Vol. 39,American Astronomical Society Meeting Abstracts, 981Jones, R. L., Yoachim, P., Chandrasekharan, S., et al. 2014,in Proc. SPIE, Vol. 9149, Observatory Operations:Strategies, Processes, and Systems V, 91490BJones, R. L., Slater, C. T., Moeyens, J., et al. 2018, Icarus,303, 181, doi: 10.1016/j.icarus.2017.11.033Joudaki, S., Blake, C., Johnson, A., et al. 2018, MNRAS,474, 4894, doi: 10.1093/mnras/stx2820Juri´c, M., Becker, A., Shaw, R., Krughoff, K. S., & Kantor,J. 2013, Winter 2013 LSST DM Data Challenge ReleaseNotes, DM Technical Note DMTN-035, LSST DataManagement. https://dmtn-035.lsst.ioJuri´c, M., Ciardi, D., & Dubois-Felsmann, G. 2017a, LSSTScience Platform Vision Document, LSST DocumentLSE-319, LSST. https://ls.st/LSE-319Juri´c, M., Ivezi´c, ˇZ., Brooks, A., et al. 2008, ApJ, 673, 864,doi: 10.1086/523619Juri´c, M., Kantor, J., Lim, K., et al. 2015, ArXiv e-prints.https://arxiv.org/abs/1512.07914Juri´c, M., et al. 2017b, LSST Data Products DefinitionDocument, LSST Document LSE-163, LSST.https://ls.st/LSE-163 Kaiser, N., Aussel, H., Burke, B. E., et al. 2002, inProc. SPIE, Vol. 4836, Survey and Other TelescopeTechnologies and Discoveries, ed. J. A. Tyson & S. Wolff,154–164Kaiser, N., Burgett, W., Chambers, K., et al. 2010, inProc. SPIE, Vol. 7733, Ground-based and AirborneTelescopes III, 77330EKalirai, J. S., Hansen, B. M. S., Kelson, D. D., et al. 2008,ApJ, 676, 594, doi: 10.1086/527028Kalirai, J. S., Richer, H. B., Fahlman, G. G., et al. 2001,AJ, 122, 257, doi: 10.1086/321140Kann, D. A., Klose, S., Zhang, B., et al. 2010, ApJ, 720,1513, doi: 10.1088/0004-637X/720/2/1513Kasliwal, M. M. 2011, PhD thesis, California Institute ofTechnologyKaspi, S., Brandt, W. N., Maoz, D., et al. 2007, ApJ, 659,997, doi: 10.1086/512094Keller, S. C., Schmidt, B. P., Bessell, M. S., et al. 2007,PASA, 24, 1, doi: 10.1071/AS07001Kilbinger, M. 2015, Reports on Progress in Physics, 78,086901, doi: 10.1088/0034-4885/78/8/086901Knezevic, Z., & Milani, A. 2005, Highlights of Astronomy,13, 758Knox, L., Song, Y.-S., & Tyson, J. A. 2006, PhRvD, 74,023512, doi: 10.1103/PhysRevD.74.023512Komatsu, E., Dunkley, J., Nolta, M. R., et al. 2009, ApJS,180, 330, doi: 10.1088/0067-0049/180/2/330Komossa, S. 2015, Journal of High Energy Astrophysics, 7,148, doi: 10.1016/j.jheap.2015.04.006Kowalski, A. F., Hawley, S. L., Holtzman, J. A.,Wisniewski, J. P., & Hilton, E. J. 2010, ApJL, 714, L98,doi: 10.1088/2041-8205/714/1/L98Kowalski, M., Rubin, D., Aldering, G., et al. 2008, ApJ,686, 749, doi: 10.1086/589937Krause, E., & Eifler, T. 2017, MNRAS, 470, 2100,doi: 10.1093/mnras/stx1261Kulkarni, S. R., Ofek, E. O., Rau, A., et al. 2007, Nature,447, 458, doi: 10.1038/nature05822Kundi´c, T., Turner, E. L., Colley, W. N., et al. 1997, ApJ,482, 75, doi: 10.1086/304147Kurucz, R. L. 1979, ApJS, 40, 1, doi: 10.1086/190589—. 1993, SYNTHE spectrum synthesis programs and linedata (Smithsonian Astrophysical Observatory)Laher, R. R., Masci, F. J., Groom, S., et al. 2017, ArXive-prints. https://arxiv.org/abs/1708.01584Laureijs, R., Amiaux, J., Arduini, S., et al. 2011, ArXive-prints. https://arxiv.org/abs/1110.3193Law, D. R., Johnston, K. V., & Majewski, S. R. 2005, ApJ,619, 807, doi: 10.1086/426779 Ivezi´c, Kahn, Tyson, Abel, Acosta, Allsman, Alonso, AlSayyad, Anderson, et al.
Law, N. M., Kulkarni, S. R., Dekany, R. G., et al. 2009,PASP, 121, 1395, doi: 10.1086/648598Leistedt, B., Peiris, H. V., & Roth, N. 2014, PhysicalReview Letters, 113, 221301,doi: 10.1103/PhysRevLett.113.221301Lenz, D. D., Newberg, J., Rosner, R., Richards, G. T., &Stoughton, C. 1998, ApJS, 119, 121, doi: 10.1086/313155L´epine, S. 2008, AJ, 135, 2177,doi: 10.1088/0004-6256/135/6/2177Levi, M., Bebek, C., Beers, T., et al. 2013, ArXiv e-prints.https://arxiv.org/abs/1308.0847Linder, E. V. 2003, PhRvD, 68, 083504,doi: 10.1103/PhysRevD.68.083504Lintott, C., Schawinski, K., Bamford, S., et al. 2011,MNRAS, 410, 166, doi: 10.1111/j.1365-2966.2010.17432.xLonghitano, M., & Binggeli, B. 2010, A&A, 509, A46,doi: 10.1051/0004-6361/200913109Lorimer, D. R., Bailes, M., McLaughlin, M. A., Narkevic,D. J., & Crawford, F. 2007, Science, 318, 777,doi: 10.1126/science.1147532Lowry, S. C., Fitzsimmons, A., Cartwright, I. M., &Williams, I. P. 1999, A&A, 349, 649LSST Dark Energy Science Collaboration. 2012, ArXive-prints. https://arxiv.org/abs/1211.0310LSST Science Collaboration, Abell, P. A., Allison, J., et al.2009, ArXiv e-prints. https://arxiv.org/abs/0912.0201LSST Science Collaboration, Marshall, P., Anguita, T.,et al. 2017, ArXiv e-prints.https://arxiv.org/abs/1708.04058Lucas, P. W., & Roche, P. F. 2000, MNRAS, 314, 858,doi: 10.1046/j.1365-8711.2000.03515.xLue, A., Scoccimarro, R., & Starkman, G. 2004, PhRvD,69, 044005, doi: 10.1103/PhysRevD.69.044005Luhman, K. L. 2014, ApJL, 786, L18,doi: 10.1088/2041-8205/786/2/L18Lupton, R., Blanton, M. R., Fekete, G., et al. 2004, PASP,116, 133, doi: 10.1086/382245Macaulay, E., Davis, T. M., Scovacricchi, D., et al. 2017,MNRAS, 467, 259, doi: 10.1093/mnras/stw3339MacLeod, C. L., Ivezi´c, ˇZ., Kochanek, C. S., et al. 2010,ApJ, 721, 1014, doi: 10.1088/0004-637X/721/2/1014MacLeod, C. L., Ross, N. P., Lawrence, A., et al. 2016,MNRAS, 457, 389, doi: 10.1093/mnras/stv2997Mahabal, A. A., Djorgovski, S. G., Drake, A. J., et al. 2011,Bulletin of the Astronomical Society of India, 39, 387.https://arxiv.org/abs/1111.0313Majewski, S. R., Skrutskie, M. F., Weinberg, M. D., &Ostheimer, J. C. 2003, ApJ, 599, 1082,doi: 10.1086/379504 Mamajek, E. E. 2012, ArXiv e-prints.https://arxiv.org/abs/1210.1616Mana, A., Giannantonio, T., Weller, J., et al. 2013,MNRAS, 434, 684, doi: 10.1093/mnras/stt1062Mandelbaum, R. 2017, ArXiv e-prints.https://arxiv.org/abs/1710.03235Mandelbaum, R., Seljak, U., Kauffmann, G., Hirata, C. M.,& Brinkmann, J. 2006, MNRAS, 368, 715,doi: 10.1111/j.1365-2966.2006.10156.xMarcus, R. A., Ragozzine, D., Murray-Clay, R. A., &Holman, M. J. 2011, ApJ, 733, 40,doi: 10.1088/0004-637X/733/1/40Martin, D. C., Fanson, J., Schiminovich, D., et al. 2005,ApJL, 619, L1, doi: 10.1086/426387Martini, P., & Schneider, D. P. 2003, ApJL, 597, L109,doi: 10.1086/379888Mathieu, R. D. 2000, in ASP Conf. Ser., Vol. 198, StellarClusters and Associations: Convection, Rotation, andDynamos, ed. R. Pallavicini, G. Micela, & S. Sciortino,517Matthews, D. J., & Newman, J. A. 2010, ApJ, 721, 456,doi: 10.1088/0004-637X/721/1/456M´enard, B., Scranton, R., Schmidt, S., et al. 2013, ArXive-prints. https://arxiv.org/abs/1303.4722Merloni, A., Predehl, P., Becker, W., et al. 2012, ArXive-prints. https://arxiv.org/abs/1209.3114Metzger, B. D. 2017, Living Reviews in Relativity, 20, 3,doi: 10.1007/s41114-017-0006-zMetzger, B. D., & Berger, E. 2012, ApJ, 746, 48,doi: 10.1088/0004-637X/746/1/48Miyazaki, S., Komiyama, Y., Kawanomoto, S., et al. 2018,PASJ, 70, S1, doi: 10.1093/pasj/psx063Mo, H. J., Jing, Y. P., & White, S. D. M. 1996, MNRAS,282, 1096, doi: 10.1093/mnras/282.3.1096Monet, D. G., Levine, S. E., Canzian, B., et al. 2003, AJ,125, 984, doi: 10.1086/345888Moniez, M. 2003, A&A, 412, 105,doi: 10.1051/0004-6361:20031478Morokuma, T., Inada, N., Oguri, M., et al. 2007, AJ, 133,214, doi: 10.1086/509701Morrison, C. B., Scranton, R., M´enard, B., et al. 2012,MNRAS, 426, 2489,doi: 10.1111/j.1365-2966.2012.21826.xMunn, J. A., Monet, D. G., Levine, S. E., et al. 2004, AJ,127, 3034, doi: 10.1086/383292Najita, J., Willman, B., Finkbeiner, D. P., et al. 2016,ArXiv e-prints. https://arxiv.org/abs/1610.01661 he Large Synoptic Survey Telescope Ivezi´c, Kahn, Tyson, Abel, Acosta, Allsman, Alonso, AlSayyad, Anderson, et al.
Ross, N. P., Myers, A. D., Sheldon, E. S., et al. 2012, ApJS,199, 3, doi: 10.1088/0067-0049/199/1/3Rossetto, B. M., Santiago, B. X., Girardi, L., et al. 2011,AJ, 141, 185, doi: 10.1088/0004-6256/141/6/185Rowe, B. T. P., Jarvis, M., Mandelbaum, R., et al. 2015,Astronomy and Computing, 10, 121,doi: 10.1016/j.ascom.2015.02.002Rowe, J. F., Bryson, S. T., Marcy, G. W., et al. 2014, ApJ,784, 45, doi: 10.1088/0004-637X/784/1/45Schlafly, E. F., Finkbeiner, D. P., Juri´c, M., et al. 2012,ApJ, 756, 158, doi: 10.1088/0004-637X/756/2/158Schlegel, D. J., Finkbeiner, D. P., & Davis, M. 1998, ApJ,500, 525, doi: 10.1086/305772Science Working Group of the LSST, & Strauss, M. A.2004, Towards a Design Reference Mission for the LargeSynoptic Survey Telescope, Document-26952, LSST.https://ls.st/Document-26952Scolnic, D., Kessler, R., Brout, D., et al. 2018, ApJ, 852,doi: 10.3847/2041-8213/aa9d82Scovacricchi, D., Nichol, R. C., Bacon, D., Sullivan, M., &Prajs, S. 2016, MNRAS, 456, 1700,doi: 10.1093/mnras/stv2752Scovacricchi, D., Nichol, R. C., Macaulay, E., & Bacon, D.2017, MNRAS, 465, 2862, doi: 10.1093/mnras/stw2878Scranton, R., Connolly, A., Krughoff, S., et al. 2007, ArXive-prints. https://arxiv.org/abs/0709.0752Seljak, U. 2009, Physical Review Letters, 102, 021302,doi: 10.1103/PhysRevLett.102.021302Seo, H.-J., & Eisenstein, D. J. 2003, ApJ, 598, 720,doi: 10.1086/379122Sesar, B., Ivezi´c, ˇZ., Lupton, R. H., et al. 2007, AJ, 134,2236, doi: 10.1086/521819Sesar, B., Ivezi´c, ˇZ., Grammer, S. H., et al. 2010, ApJ, 708,717, doi: 10.1088/0004-637X/708/1/717Shara, M. M. 2006, AJ, 131, 2980, doi: 10.1086/501502Sharma, S., Johnston, K. V., Majewski, S. R., Bullock, J.,& Mu˜noz, R. R. 2011, ApJ, 728, 106,doi: 10.1088/0004-637X/728/2/106Shen, K. J., & Bildsten, L. 2009, ApJ, 692, 324,doi: 10.1088/0004-637X/692/1/324Shen, Y., Brandt, W. N., Dawson, K. S., et al. 2015, ApJS,216, 4, doi: 10.1088/0067-0049/216/1/4Sheppard, S. S., & Trujillo, C. 2016, AJ, 152, 221,doi: 10.3847/1538-3881/152/6/221Shields, G. A., & Bonning, E. W. 2008, ApJ, 682, 758,doi: 10.1086/589427Silsbee, K., & Tremaine, S. 2016, AJ, 152, 103,doi: 10.3847/0004-6256/152/4/103Skrutskie, M. F., Cutri, R. M., Stiening, R., et al. 2006, AJ,131, 1163, doi: 10.1086/498708 Smartt, S. J. 2009, ARA&A, 47, 63,doi: 10.1146/annurev-astro-082708-101737Smith, N. 2014, ARA&A, 52, 487,doi: 10.1146/annurev-astro-081913-040025Smith, N., Li, W., Silverman, J. M., Ganeshalingam, M., &Filippenko, A. V. 2011, MNRAS, 415, 773,doi: 10.1111/j.1365-2966.2011.18763.xSolontoi, M. 2010, PhD thesis, University of WashingtonSong, Y.-S., & Knox, L. 2004, PhRvD, 70, 063510,doi: 10.1103/PhysRevD.70.063510Spergel, D., Gehrels, N., Baltay, C., et al. 2015, ArXive-prints. https://arxiv.org/abs/1503.03757Staniszewski, Z., Ade, P. A. R., Aird, K. A., et al. 2009,ApJ, 701, 32, doi: 10.1088/0004-637X/701/1/32Stassun, K., Paegert, M., De Lee, N. M., & Cargile, P.2013, in American Astronomical Society MeetingAbstracts, Vol. 221, American Astronomical SocietyMeeting Abstracts he Large Synoptic Survey Telescope57