Transients and alerts
with Gaia
Łukasz Wyrzykowski
(pron. woocash vizhikovski)
(or just lucas)
with
Simon Hodgkin, Ross Burgon
and
Gerry Gilmore, Nick Walton,
Vasily Belokurov, Wyn Evans
Institute of Astronomy, University of Cambridge, UK
email: wyrzykow@ast.cam.ac.uk
Next Generation Sky Surveys : The Transient Sky
eSI, QUB, Belfast, 20 January 2010
Gaia: Complete, Faint, Accurate
Łukasz Wyrzykowski, IoA Cambridge UK
Gaia Science Alerts
Satellite and System
ESA-only mission
Launch date: spring 2012
Lifetime: 5 years (1 year potential extension)
Launcher: Soyuz–Fregat from CSG
Orbit: L2 Lissajous orbit
Ground station: Cebreros and New Norcia
Downlink rate: 4–8 Mbps
• Mass: 2120 kg (payload 743 kg)
• Power: 1631 W (payload 815 W)
Figures courtesy EADS-Astrium
Łukasz Wyrzykowski, IoA Cambridge UK
Gaia Science Alerts
Scanning law
•
•
•
•
Two telescopes - time between subsequent FOVs: 106.5m
Time between successive scans: 6 h
Field revisited every ~70 days
Each object measured ~80 times (200 at the nodes)
Lukasz Wyrzykowski, IoA Cambridge UK
Gaia Science Alerts
Gaia’s Focal Plane
•
•
Chip transit: 4.4s
Field transit (9 astrometric CCDs): 40s
Lukasz Wyrzykowski, IoA Cambridge UK
Gaia Science Alerts
Science Alerts
aims:
•
detect unexpected and rapid changes in the flux
•
or appearance of new objects
•
trigger ground-based follow-up
•
provide targets to the community to be studied at peculiar states
methods:
•
run in near-real-time: between couple of hours and 24h after
observation
•
use photometric data, calibrated roughly
•
release of an alert through VO
Łukasz Wyrzykowski, IoA Cambridge UK
Gaia Science Alerts
Aims
Łukasz Wyrzykowski, IoA Cambridge UK
Gaia Science Alerts
GRBs optical counterparts
Potential Triggers
R Coronae Borealis
FU Orionis and similar
M-dwarf flares
Be stars
Microlensing events
Supernovae
Asteroids
Dwarf novae
NEW THINGS??
Classical novae
Łukasz Wyrzykowski, IoA Cambridge UK
Lensed supernovae
Gaia Science Alerts
Rates of alerts (preliminary)
contaminants(?)
interesting
alerting object
5-yrs
(Entire Mission)
main location
Supernovae <19 mag
6000
out of plane
Microlensing (bulge)
~1000
bulge/plane
Microlensing (all sky)
~700
out of plane
GRB optical counterparts
~hundreds (?)
out of plane
R CrB-type stars
~hundreds (?)
gal. plane
CN
150
gal. plane
FU Ori
14
gal. plane
Eclipsing binaries
a million (?)
gal. plane
AGNs
500,000 (?)
out of plane
Asteroids
thousands (?)
out of plane
Be stars
thousands (?)
gal. plane
Long period variables/Miras
thousands (?)
gal. plane
M-dwarf flares
2000
gal. plane
DN (U Gem) (except rare big flares)
500 (?)
gal. plane
Łukasz Wyrzykowski, IoA Cambridge UK
Gaia Science Alerts
Supernovae
•
•
6000 SNe to G=19
Around 1/3 before maximum
• Successive transits
will measure
consistency and
slope.
• Host galaxy
40s to cross the FOV
106.5 mins between FOVs
Field revisited after 6 hours
Łukasz Wyrzykowski, IoA Cambridge UK
contribution
determines whether
source is new to
Gaia.
Gaia Science Alerts
Microlensing Events
• >3000 events expected to occur towards the bulge,
but many lost due to crowding (exception: Baade’s Window)
• ~ 700 events expected over all sky
• photometric alerts on 1000+ events
• ~100% long events detected - the most interesting ones
(nearby or massive lens)
time scale = 100 days
time scale = 15 days
Gaia microlensing
detection efficiency
on the rise
OGLE-III microlensing events
as seen with Gaia
efficiency
0
0.2
0.4
0.6
0.8
Figure 1: OGLE’s microlensing events as would have been observed b
indicate the data points taken during the actual event. Shown are events wi
points during the event more than 20.
When alerted in time and followed-up:
★ luminous and dark mass distribution in the Galaxy
★ dark matter in compact objects
★ extra-solar planets
Technical Note
Łukasz Wyrzykowski, IoA Cambridge UK
Gaia Science Alerts
R CrB-type stars
• very few known (~50), but ~3000 expected
• can drop up to 8 mag in brightness over a week! - ideal time-scale for Gaia sampling
• mechanism of these dimmings remains unknown
• most found in the Bulge and MCs by microlensing surveys EROS, MACHO, OGLE
• ASAS (all sky) provided some new findings, limited by 14mag
• Science Alerts will easily find new much fainter RCrB
• spectroscopic follow-up during the event can help solving their mystery
Łukasz Wyrzykowski, IoA Cambridge UK
Gaia Science Alerts
Other interesting triggers
✦ Classical and recurrent novae
- potential distance indicators
- can be detected in MW and other galaxies
- large amplitudes, wide range of time-scales
✦ FU Orionis/EX Lupi
- unstable pre-MS stars
- rare class (few known)
- several magnitudes up
- X-ray variability
- long time-scales
- FU Ori repeats every ~40 years!
✦ Gravitationally lensed distant supernovae
- unique uniform all-sky monitoring
- high redshift SNe rates, distances, H0
Łukasz Wyrzykowski, IoA Cambridge UK
Gaia Science Alerts
Common and not interesting(?)
➡ Asteroids - loads!
- faster ones can be removed after second FOV transit
- predominantly known asteroids
- the orbits will be calculated and new objects will be cross-matched
➡ Be stars - low amplitudes, blue colours
➡ Dwarf Novae - low amplitudes, repeating
➡ and many more...
Łukasz Wyrzykowski, IoA Cambridge UK
Gaia Science Alerts
Monitoring System
(Watch List)
‣ monitoring of known anomalous objects
‣ easily coded - just add an object to the list of alerts
‣ many of the things we alert on will be subsequently monitored
‣ but some will not alert (e.g. highly variable), but are known to be
anomalous
‣ for example: known RCrB stars can be monitored
in order to alert on a dimming event
‣ TBC if in agreement with the Gaia data policy
Methods
Łukasz Wyrzykowski, IoA Cambridge UK
Gaia Science Alerts
Science Alerts
operation scheme
New
Data
Photom.
processing within 1 day and are stored
• Detection is run on data calibrated with the
most recent calibration available
Data Ingestion
Database
• Numerous detection algorithms are used
Calibs.
Photom.
Flux Alert
Detector
Candidate
Alerts
Curator
• SDSS
• 2MASS
• NED
• ASAS
• OGLE
• New Data arrive to Cambridge after basic pre-
Asteroids
Gaia data
Classifier
Catalogues
Cross-match
Classifier
+ Gaia pre-launch
catalogue
Publisher
Łukasz Wyrzykowski, IoA Cambridge UK
Detection
Algorithms
• Candidate alerts are classified using Gaia data
• Sources are cross-matched with available
catalogues and classified further
• Cross-matched asteroids are removed
• Alerts are stored on the Server and released
Gaia
Alerts
Server
VOEvent
Gaia Science Alerts
Detection: the simplest algorithm
Mean
History
threshold
History
False alert
Łukasz Wyrzykowski, IoA Cambridge UK
Gaia Science Alerts
Curation - Gaia Classification:
data used in Decision Tree
- G-band photometry (light curve)
- BP/RP colour
- raw BP and RP spectra (classification using Self-Organizing Maps)
- morphology of the source (galaxy/star)
- source motion flags (fast asteroid?)
- Gaia catalogues (later in the mission)
- galaxies
- variable stars classified into types
- astrophysical parameters, e.g. Teff, spectral type
Łukasz Wyrzykowski, IoA Cambridge UK
Gaia Science Alerts
Curation - Cross-matching
uses existing catalogues available through Virtual Observatory,
Astrogrid or local copies,
e.g. SDSS, 2MASS, NED, ASAS, OGLE
- star, source close to galaxy or orphan new source?
- magnitudes and colours in optical and IR
- X-ray source, gamma source?
- time-domain photometry
- variability classification (e.g. recurrent nova, eclipsing)
- asteroids flagged by other Gaia units
- anomalies alerted by other surveys (e.g. GRBs alerts)
Łukasz Wyrzykowski, IoA Cambridge UK
Gaia Science Alerts
Outreach with Science Alerts
๏
Science Alerts can be exciting not only for professionals
๏
a platform (EU RTN) will be build to provide alerts to the general public
๏
not only alerts, but also tools for observing, data reduction, studying
๏
cool projects, e.g. “Adopt a Supernovae”
๏
amateurs, schools, universities...
Łukasz Wyrzykowski, IoA Cambridge UK
Gaia Science Alerts
Challenges and requirements
๏
Data storage and access
Science Alerts will independently store all raw observations from the entire mission. This is about
1012 measurements (~20TB). Storing and access must be as quick as possible.
๏
Processing speed
Also data processing must be quick. About 107 objects will be checked every day. Parallel processing
needed.
๏
Design of detection algorithms and classification of alerts
Remove spurious detections (most of the alerts!), classify, attach a priority
๏
Fine-tuning of the detection and classification procedures
In the first 3 months of the mission the alerts will not be released and will be internally verified in
order to fine-tune the procedures. After that still some minor fine-tuning will be required.
๏
Well organised follow-up network
Gaia is a survey-only telescope with relatively poor sampling. Without a follow-up its alerts, in most
cases, will be meaningless.
Łukasz Wyrzykowski, IoA Cambridge UK
Gaia Science Alerts
http://www.ast.cam.ac.uk/research/gsawg/
Łukasz Wyrzykowski, IoA Cambridge UK
Gaia Science Alerts
Preliminary Advert
Science from Gaia Alerts
workshop supported by GREAT
(Gaia Research for European Astronomy Training)
Cambridge, 23-25 June 2010
more info soon on Gaia Science Alert WG wiki:
http://www.ast.cam.ac.uk/research/gsawg/
Łukasz Wyrzykowski, IoA Cambridge UK
Gaia Science Alerts