AN ANTAR ARES: Making aking real eal-ti time astr trono nomy - - PowerPoint PPT Presentation

AN ANTAR ARES: Making aking real eal-ti time astr trono nomy y with th LS LSST possible

MO MONIKA NIKA D. SO SORAISAM

NATIONAL OPTICAL ASTRONOMY OBSERVATORY LSST2018 PROJECT AND COMMUNITY WORKSHOP, AUG 16

ARIZONA-NOAO TEMPORAL ANALYSIS & RESPONSE to EVENTS SYSTEM

NOAO: Abi Saha, Tom Matheson, Chien-Hsiu Lee, Monika Soraisam (soraisam@noao.edu), Carl Stubens (cstubens@noao.edu), Adam Scott, Glenn Eychanner STScI: Gautham Narayan (Lasker Fellow; gnarayan@stsci.edu) University of Arizona: Rick Snodgrass, John Kececioglu, Carlos Scheidegger Grads & Undergrads: Zhe Wang, Zhenge Zhao, Eric Welch, Navdeep Singh, Eric

• M. Evans, Songzhe Zhu, Daniel Muthukrishna (Cambridge)

Time Domain Astronomy (TDA) – then, now

• TDA has been historically driven by visual/human

inspection to find a particular class of objects

• Current wide-field surveys rely on it as well

LSST2018: Following-up LSST Alerts, Aug 16

Issues: prone to errors, irreproducible, limited by data rate, …

Time Domain Astronomy (TDA) – then, now, near future

• TDA has been historically driven by visual/human

inspection to find a particular class of objects

• Current wide-field surveys rely on it as well
• LSST:
• colossal volume of data, an astronomic alert

rate (≈ 10 million/night, 1 image/40 secs)

• Community’s diverse follow-up interests, e.g.,

rare/novel events, known unknowns,…

• Rare short-lived alerts necessitate prompt

follow-up

LSST2018: Following-up LSST Alerts, Aug 16

Issues: prone to errors, irreproducible, limited by data rate, … Follow-up resources available to the entire astronomical community no match for the LSST alert rate

Time Domain Ecosystem

• Alert generators: difference imaging,

real/bogus classification, moving

• bject assessment
• Brokers: manage the alert flow,

winnowing down to the most interesting lot Ø characterize alerts automatically Ø distribute to interested parties Ø (can) interface with TOMs to trigger rapid follow-up of transient alerts

LSST2018: Following-up LSST Alerts, Aug 16

LSST

Initial Ingest Existing Catalogs 1,000 images / night Thumbnails Updated Calibration Data daily Camera Alerts Up to 10,000 alerts / image quarterly Measurement Recalibration AstroObject & Alert Association AstroObject- Associated Alerts Feature Evaluation Extended Alerts Alert Characterization Characterized Alerts Initial Filtering Level I Alerts Advanced Feature Evaluation, Characterization and Filtering Level II Alerts 1,000 / image 100 / image 1.0 / image Taxonomy of Characterizations (curated) diverted alerts diverted alerts Filtering Trees (curated) Characterization Algorithms (curated) Locus-Aggregated Annotated Alerts Periodic Queries Derive Features New AstroObject Identification Position Estimation Moving Objects (curated) Moving Object Identification Feature Algorithms (curated) Locus Aggregation

ANTARES

Theoretical Predictions Algorithm Vetting Professionals, citizen scientists forking forking start at image acquisition Other External Alerts External Alert Ingestion Theorists Alert Throttling Thumbnail Separation Combo Assembly Novel Alerts 0.1 / image

“Rarest of the Rare”

Annotated Coalesced Alerts External Alert Brokers

Researchers and public

Ultra Characterization & Filtering 23 Nov 2016 Alert Replica Coalescing Alert Routing black = us green = not us blue = catalog by-products red = alert flow Annotated Coalesced Alert Rarest Alerts Aggregated AstroObject Catalogs (curated) Aggregated AstroObject Catalogs (curated) Aggregated AstroObject Catalogs (curated) forking Combo Replicas Selected Alert Replicas Combo Feature Evaluation Advanced Feature Evaluation Combo Coalescing Touchstone (curated) Touchstone (curated) Touchstone (curated) Alert Simulator 100 / night

INGEST ASSOCIATE FEATURIZE FILTER FILTER FILTER NOVEL ALERTS

ANTARES DOES

• ingest alerts generated outside

ANTARES

• annotate alerts with their history

and external contextual data from associated astro-object

• derive features for the alert &

compare against the ‘Touchstone’

• filter/rank the alerts
• distribute novel alerts, while the rest

are stored with their added values in the LADB

LSST2018: Following-up LSST Alerts, Aug 16

ANTARES DOES NOT

• perform difference imaging

and identification of sources

• n images
• real-bogus assessment of

the source

• moving object identification
• follow-up coordination

ANTARES STATUS

• Functional prototype being hosted at the cluster at Arizona UITS
• Basic architecture design implemented, robust system API
• Astro-object catalogs covering wide range of wavelengths (2MASS,

ALLWISE, NED, Chandra, Gaia DR1, Galex, SDSS, etc.)

• Improved front-ends: web-based front-end Alert Viewer, including data

provenance support (DAFCAA)

• w.r.t ZTF MSIP alert stream, ANTARES can:
• connect to ZTF Kafka stream, ingest ZTF .avro formatted data
• run alert processing pipeline on ZTF alerts
• detect interesting objects automatically

LSST2018: Following-up LSST Alerts, Aug 16

LSST2018: Following-up LSST Alerts, Aug 16

LSST2018: Following-up LSST Alerts, Aug 16

Data provenance

LSST2018: Following-up LSST Alerts, Aug 16

Data provenance

LSST2018: Following-up LSST Alerts, Aug 16

ANTARES API

• ANTARES will accept custom

filters from the community.

• Filters determine which alerts

are distributed in real-time to downstream consumers.

• The ANTARES API will define

how astronomers will write their filters and what data they have access to. (API documentation coming soon…)

LSST2018: Following-up LSST Alerts, Aug 16

LSST2018: Following-up LSST Alerts, Aug 16

SUMMARY

• ANTARES strives to serve as a community alert broker for large-scale optical time-

domain surveys (LSST, ZTF).

• Community involvement will be a big factor for success of brokering
• or brokering of success for time-domain efforts.

LSST2018: Following-up LSST Alerts, Aug 16