overview & status of lsst...2018/07/10 · overview of lsst the deep, wide, fast survey lsst...

Overview & Status of LSSTJohn Swinbank • [email protected]

Deputy Project Manager, LSST Data Management

Overview of LSST

Not “just” a telescope

�2LSST Solar System Readiness Sprint • 2018-07-10 • University of Washington

• An integrated survey system: ‣ A wide field survey telescope ‣ The “world’s biggest digital camera” ‣ Automatic scheduling & operations

• A decade-long survey • Fully reduced data products • An interactive data access & analysis environment

Overview of LSST

The Deep, Wide, Fast Survey


• Starting 2022

• Running for a 10 years

• 18000+ deg2

• 30s exposure time per visit

• ~825 visits per point

- ~50 to ~200 per filter

• rAB~24.5/visit; rAB~27.5 total

• Detailed survey strategy still being developed

19

Fig. 18.— The distribution of the r band visits on the sky for asimulated realization of the baseline cadence. The sky is shown inthe equal-area Mollweide projection in equatorial coordinates (thevernal equinoctial point is in the center, and the right ascension isincreasing from right to left). The number of visits for a 10-yearsurvey, normalized to the SRD design value of 184, is color-codedaccording to the legend. The two regions with smaller number ofvisits than the main survey (“mini-surveys”) are the Galactic plane(arc on the right) and the region around the South Celestial Pole(bottom). The so-called “northern Ecliptic region” (upper left)has received more visits than the main survey in this particularsimulation (in order to increase completeness for moving objectsby increasing the coverage of the Ecliptic plane). Deep drillingfields, with a much higher number of visits than the main survey,are also visible as small circles. The fields were dithered on sub-field scales and pixels with angular resolution of ⇠30 arcmin wereused to evaluate and display the coverage.

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 (⇠1minute), and observations of “special” regions such asthe Ecliptic, Galactic plane, and the Large and SmallMagellanic Clouds. We are also considering a third typeof survey, micro-surveys, that would use about 1% of thetime (which still represents 25 nights on a unique 8m-class telescope).

3.1.1. The Main Deep-Wide-Fast Survey

The observing strategy for the main survey will be opti-mized for the homogeneity of depth and number of visits.In times of good seeing and at low airmass, preference isgiven to r-band and i-band observations. As often aspossible, each field will be observed twice, with visitsseparated by 15-60 minutes. This strategy will providemotion vectors to link detections of moving objects inthe Solar System, and fine-time sampling for measuringshort-period variability. The ranking criteria also ensurethat the visits to each field are widely distributed in po-sition angle on the sky and rotation angle of the camerain order to minimize systematic e↵ects in galaxy shapedetermination.The universal cadence provides most of LSST’s power

for detecting Near Earth Objects (NEO) and naturallyincorporates the southern half of the ecliptic within its18,000 square degrees (the northern half lies above thedesired airmass limits, X . 1.5). NEO sample complete-ness for the smallest bodies (⇠140m in diameter, per theCongressional NEO mandate) is greatly enhanced, how-ever, by the addition of a crescent on the sky within 10

degrees of the northern ecliptic (see Fig. 18). 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 deg2 around the SouthCelestial Pole, allowing coverage of the Large and SmallMagellanic Clouds.Finally, the universal cadence proposal excludes obser-

vations in a region of 1,000 square degrees around theGalactic Center, where the high stellar density leads to aconfusion limit at much brighter magnitudes than thoseattained in the rest of the survey. Within this region,the Galactic Center proposal provides 30 observations ineach of the six filters, distributed roughly logarithmicallyin time (it may not be necessary to use the u and g filtersfor this heavily extincted region).The resulting sky coverage for the LSST baseline ca-

dence, based on detailed operations simulations, is shownfor the r band in Fig. 18. The anticipated total numberof visits for a ten-year LSST survey is about 2.8 million(⇠5.6 million 15-second long exposures). The per-bandallocation of these visits is shown in Table 1.

3.1.2. Mini-surveys

Although the uniform treatment of the sky provided bythe universal cadence proposal can satisfy the majorityof LSST scientific goals, roughly 10% of the time will beallocated to other strategies that significantly enhancethe scientific return. These surveys aim to extend theparameter space accessible to the main survey by goingdeeper or by employing di↵erent time/filter sampling.As an example of a mini-survey, consider a program

that uses one hour of observing time per night to observea relatively small region of sky to substantially greaterdepth in individual visits. Accounting for read-out timeand filter changes, it could obtain about 50 consecutive15-second exposures in each of four filters in an hour.If a field is visited every two days over four months,about 600 deg2 can be observed with this cadence over10 years. Taking weather into account, the selected fieldswould each have on average about 40 hour-long sequencesof 200 exposures each. Each observation in a sequencewould have an equivalent 5-� depth of r ⇠ 24.5, and eachfilter subsequence when coadded would be 2 magnitudesdeeper than the main survey visits (r ⇠ 26.5). Whenall 40 sequences and the main survey visits are coadded,they would extend the depth to r ⇠ 28.This data set would be excellent for a wide variety of

science programs. The individual sequences would besensitive to 1% variability on hourly time scales, allow-ing discovery of planetary eclipses. If these fields wereselected at Galactic latitudes of |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 ⇠ 26.5(assuming observations in 4 bands, every 2 days over 120days, and accounting for weather losses), and thus en-able deep searches for SN, trans-Neptunian objects, andother faint transient, moving and variable sources. Forexample, the SN sample would be extended to redshifts ofz ⇠ 1.2, with more densely sampled light curves than ob-tained from the universal cadence. Such sequences wouldalso serve as excellent tests of our photometric calibra-

Ivezic et al, arXiv:0805.2366

“Deep Drilling” fieldsMini-surveysNorthern Ecliptic region

Telescope & Site

Progress on the summit


Rendering 2012

Photo May 2018• Summit facility building completed

March 2018

• Dome due late this year

Telescope & Site

Telescope Mount Assembly


• 300 ton moving structure

• 10 deg / second rotation

• 10 deg / sec2 acceleration

• Under construction by Asturfeito, S.A., Spain

• Assembly on the mountain in late 2018

• Photo: April 2018

Telescope & Site

Mirrors


• M1M3 polished in 2014

• Integration activities ongoing in Tucson

• On summit mid-2019

• M2 mirror & assembly on track for shipping from vendor October 2018

M2

M1M3

Telescope & Site

Auxiliary Telescope


• 1.2 m telescope, in its own dome next to LSST

• Spectrophotometric measurements of starsto probe atmospheric absorption in support of LSST calibration

• Telescope now on site; spectrograph in the lab in Tucson

• First light expected early 2019

Camera

Basic Parameters


• 3.2 Gigapixels

• ~7.2 GB per exposure

• 2 second readout

• 0.2 arcsec pixels

• 1.65 by 3 metres; 2800kg

• Size of a small car

SLAC National Accelerator Laboratory

Camera

Focal Plane


• 63 cm diameter

• 189 sensors packed into21 “rafts” of 9 sensors each

• 4k by 4k pixel sensors

• 2 second readout time

• Cooled to 173 K

• Raft Electronics Board (REB) make each raft a ~150 MPix camera

Camera

Sensors & Rafts


• Sensors fabricated by two vendors

• Procurement almost complete

• 242 sensors accepted (including reserves)

• 10 rafts assembled and accepted

• All science rafts expected to be complete in January 2019

e2vsensors

ITL sensors

Camera

Optics


• L1 & L2 lenses polished and accepted for coating

• L3 polishing in progress; coatingexpected mid-2018

• Filter fabrication and coating contracts in place;filters currently in production.

L3L2 L1

L1 inspection L2 at coating vendor

Camera

Commissioning Camera


• Equivalent to a single raft of the full LSST camera

• Used in system integration & early commissioning activities

• Delivery on track for Jan2019; on summit mid-year

ComCam Dewar

ComCam raftassembly at BNL

Data Management

The DM System


DataReleaseDataProductsviaannualdatareleases

11DataReleasesin10yearsFinaldatabasecatalog:15PB

PromptDataProductsvianightlyalertstreams

Average10millionalertspernightIssuedwithin60 sofshutterclose

20TBrawdatapernight(withcalibrationexposures)

Portal! Notebooks!

User Databases!

LSST Science Platform!

Software Tools!User Computing!User Files!Data Releases! Alert Streams!

Web APIs!

Internet!

LSST Users!LSSTSciencePlatform

Dataaccess&end-usercomputing

Current&previousdatareleases

Alertdatabase&“mini-broker”

Data Management

Data Products


• A stream of ~10 million time-domain events per night, detected and transmitted to event distribution networks within 60 seconds of observation.

• A catalog of orbits for ~6 million bodies in the Solar System.

• A catalog of ~37 billion objects (20B galaxies, 17B stars), ~7 trillion observations (“sources”), and ~30 trillion measurements (“forced sources”), produced annually, accessible through online databases.

• Deep co-added images.

• Services and computing resources at the Data Access Centres to enable user-specified custom processing and analysis.

• Software and APIs enabling development of analysis codes.

Data

Release

Prompt

User

Generated

Data Management

Complex Image Processing


Imag

e: R

ober

t Lup

ton

• SDSS image of the COSMOS field from Lupton et al (2004)

• ~ 3.5’

Data Management

Complex Image Processing


Imag

e: H

SC

Col

labo

ratio

n &

Rob

ert L

upto

n

• Hyper Suprime-Cam data

• PSF matched coadd

• 1.5 hours in g, r

• 3 hours in i

• Reaching approx 10 year LSST depth

• Image processing performed with prototype of LSST codebase

Data Management

Pipelines


Refer to LSE-163 fordetails of what all

these products are. https://ls.st/lse-163

Data Management

Prompt Data Processing


1. Incoming data is reduced to a Processed Visit Image (CR rejection, calibration, etc).

2. PVI is differenced against a deep template and DIASources are detected on the diffim.

3. Flux and shape of each DIASource are measured on the diffim, and PSF photometry on the PVI for total flux.

4. DIASource is matched against known DIAObjects and SSObjects.

5. If the DIASource corresponds to a known SSObject, an alert is issued at once (incl. details of the associated SSObject) and further processing proceeds in daytime.

6. Otherwise, the new DIASource is used to update the corresponding object record, or a new object record created, and an alert issued incl. details of the associated object.

7. Forced photometry is performed on the position of all new DIAObjects within 30 days.

Leanne's t

alk, next,

for details

of the

alert stre

am.

Data Management

Moving Object Processing


• During the day, we:

• Recompute orbits of known SSObjects;

• Refine association of DIASources with SSObjects;

• Search for new SSObjects based on unassociated DIASources. Mario

's talk

at 11:00

for details

of MOPS.

Data Management

Status of Pipelines


• None of the Science Pipelines are yet complete... but many parts of the system are quite usable.

• Primitives and Algorithms

‣ A rich collection of high-performance tools for working with astronomical data you can pick up and use today.

• Data Release Processing

‣ Pipelines regularly being used to reprocess Hyper Suprime-Cam data (and make their data releases).

• Alert Production

‣ Currently running small-scale tests on DECam data; rapidly scaling over the rest of this year.

• Moving Objects

‣ Solid basis of algorithmic research completed; development of code to be deployed for LSST gearing up this year.

Data Management

Processing and Data Access


• Petascale computing facility under construction at NCSA (Illinois) to host the bulk of LSST computing and data access

• Satellite system at CC-IN2p3 (France)

• Science Platform based on Firefly, JupyterLab, IVOA protocols now becoming a reality

• Now being used by Commissioning Team

Education & Public Outreach

The EPO System


• The mission of LSST's EPO is to provide non-specialists with access to a subset of LSST data through accessible & engaging online experiences, so that anyone can explore the universe and be part of the discovery process.

• Audiences:

- Formal educators teaching astronomy content at the advanced middle school, high school or college level.

- Citizen science principal investigators.

- Content developers at informal science centers& planetariums.

- Science-interested teens & adults.

Overall Status

“Earned Value” summary


Based on April 2018 Data

Fractional Completion Telescope & Site 72%Camera 81%

Data Management 40%EPO 24%

Commissioning 24%

Overall Status

Timeline to full operations


Conclusions

LSST construction is on track


• All subsystems are making rapid progress

• The DM team already has software capable of doing great science

• This is a super exciting time to be involved with LSST!

• Best wishes for a productive & enjoyable week!

overview & status of lsst...2018/07/10 · overview of lsst the deep, wide, fast survey lsst...

Documents