Follow-up of LIGO-Virgo observations of gravitational waves

1. Follow-up of LIGO-Virgo observations of gravitational waves Roy Williams (Caltech)Peter Shawhan (U. of Maryland / JSI)for the LIGO Scientific Collaborationand Virgo Collaboration Hot-Wiring the Transient UniverseNovember 14, 2013 LIGO-G1301197

2. Advanced Gravitational Wave Detectors • Factor 10 in strain means factor 1000 in volume. • Long-planned upgrades of initial GW detectors: Advanced LIGO (shown) and Advanced Virgo

3. … Now Being Installed and Tested !

4. GEO-HF Advanced Detector Network– Under Construction Advanced LIGO 600 m 4 km 3 km 3 km 4 km 4 km Advanced LIGO (pending) Advanced VIRGO

5. Credit: AEI, CCT, LSU Credit: Chandra X-ray Observatory Searches for GW Transient Sources • GWdata streams are analyzed jointly • Initially LIGO Hanford+Livingston and Virgo; later others too • Two main types of transient searches: Compact Binary Coalescence (CBC) Known waveform Matched filtering Templates for a range of component masses(spin affects waveforms too, but not so important for initial detection) Unmodelled GW Burst(< ~1 sec duration) e.g. from stellar core collapse Arbitrary waveform  Excess power Require coherent signals in detectors,using direction-dependent antenna response

6. Projecting Advanced LIGO Sensitivity Progression 2015-2018 From “Prospects for Localization of Gravitational Wave Transients by the Advanced LIGO and Advanced Virgo Observatories” arXiv:1304.0670 2017-18: ~ 9 month run, 120 - 170 Mpc Multiple Detections 2016-17: ~ 6 month run, 80 - 120 Mpc Likely Detection 2015: ~ 3 month run, 40 - 80 Mpc Possible Detection 2010: Initial LIGO, 15 Mpc

7. Low-latency GW Event Candidates • Multiple analysis pipelines running as data is collected • Generate triggers from apparent transients in the data • Estimate significance by comparing with backgrounddistribution from time-shifted analysis GEO 600 LIGO Hanford KAGRA Virgo Send infoto observers Transfer data LIGO Livingston Swift: NASA E/PO, Sonoma State U., AuroreSimonnet LIGO-India Trigger database Validate Analyze data, identify triggers,infer sky position GW data Select event candidates Estimate background

8. EM and GW • WHY Multimessenger? • Brilliant science from optical counterparts of GRBs (1998) • Brilliant science from neutrino counterparts of Supernova 1987A • EM  GW: Short GRBs and galactic SN • Will trigger a search for GW counterpart, short/long latency • SWIFT, Fermi, SNEWS • GW EM: Rapid alerts to telescope partners • Wide area search for counterpart (cf Singer talk) • Could be quite faint and red (cfNissanke talk)

9. What info from GW observatory • Time of the GW candidate • At Earth, with precision of order ~10 ms(direction-dependent) • Significance of the candidate • Expressed as an effective false alarm rate (FAR) • Sky position probability map • HEALPix grid in FITS file • Maximum Distance • Although the source could be much closer • Expect to distribute GW alerts as VOEvents over(initially) private GCN/TAN, VOEventNet and/or Skyalert

10. Alert Latency and Notes • Typical latency to generate triggers with sky position info:3 to 6 minutes • Additional time needed for validation: not yet known, but maybe ~20 min • We might provide multiple skymapswith different assumptions about the source • CBC orbit inclination: unconstrained, or face-on • GW burst polarization: unconstrained, linear, or elliptical • We may send updated information about an event after the initial alert • e.g. refined skymap or significance estimate • Will send a VOEvent referencing the first one, incrementing the version number in the IVORN

11. Position Reconstruction Accuracy vs. Time Face-on BNS 80 MpcHLV 2016-17 ~8 % contained in 20 deg2 Face-on BNS 80 MpcHLV 2017-18 ~10 % contained in 20 deg2 Face-on BNS 160 MpcHILV 2022+ ~50 % contained in 20 deg2 Face-on BNS 160 Mpc HLV 2019+ ~30 % contained in 20 deg2

12. Simulated skymaps • Posterior probability skymaps are obtained from coherent analysis • Fast coherent position reconst. (~min)Full parameter estimation can be done, but currently is slow (~days), effort underway to reduce to ~hour) Examples of BAYESTAR skymaps from simulated signalsL. Price, L. Singer, et al On gamma sky (Fermi) On Mellinger sky

13. Simulated Skymaps • Error region geometry can be non-trivial • Banana shape and/or • Disconnected islands – especially for narrowband GW burst candidates • No single "sky position" quoted in alerts to observers

14. LIGO-Virgo Approach to Partnerships • Basic Concept: Share triggers with observers who sign MOUs • Confident detection of first few GW signals will require time and care—need to avoid misinformation/rumors/media circus • Partners who sign standard MOUs will receive GW triggers promptly, exchange results, decide their own observations, analyze own data • Any apparent counterpart, may not be published or presented prior to the announcement by LIGO/VIRGO • After 4 GW events have been published, further high-confidence events will be released promptly to the public • Have sought advice and input from the astronomical community • Solicited “letters of interest” – received about 60 • Held meetings in Amsterdam (late August) and Chicago (mid September) • Very encouraging show of interest! • Feedback and suggestions from meetings have shaped some of the plans for the program

15. Summary • Gravitational wave detectors operate as a global network • Data combined and analyzed coherently • Japan and/or India critical for good localization • Advanced LIGO and Virgo upgrades are well underway • First science run is only about 2 years away • Sensitivity and position reconstruction accuracy will improve over time • EM follow-ups are exciting • Whether for the first few GW detections, or later on • Now engaging follow-up observers. Want to get involved? Write lsc-spokesperson@ligo.org(Gabriela González), and/or virgo-spokesperson@ego-gw.it(Jean-Yves Vinet).

16. Extra Slides…

17. Observing Partners During 2009–2010 • Mostly (but not all) robotic wide-field optical telescopes • Many of them used for following up GRBs, surveying for supernovae and other optical transients XRT UVOT 1.2 m 2 m 1.3 m 1 m 1 m APERTURE

18. Observing Partners During 2009–2010 25 sqdeg • Mostly (but not all) robotic wide-field optical telescopes • Many of them used for following up GRBs, surveying for supernovae and other optical transients 20×20° 3.4 sqdeg XRT UVOT 3.4 sqdeg 7.3 sqdeg 3.4 sqdeg 5.7 sqdeg 9.4 sqdeg 3.4 sqdeg 3.4 sqdeg 3.4 sqdeg FIELD OF VIEW

19. First Implementation: 2009–2010 • Analyzed 3-detector data during 10 weeks of LIGO-Virgo running • Low-latency CBC and Burst searches • Sent alerts to ten follow-up partners • Selected (RA,Dec) points, targeting nearby galaxies in GW error region • Various communication protocols and feedback methods • General description: A&A 539, A124 • CBC low-latency analysis: A&A 541, A155 • Eight event candidates were followed up by at least one telescope • No stand-out candidates, unfortunately • Swift (XRT/UVOT) image analysis results: Evans et al., ApJS203, 28 • Ground-basedtelescopeanalysisresults: Aasi et al., arXiv:1304.0670 • Lessons learned: standardize, give observers more freedom +VLA (later)

20. Possible VOEvent content for a GW alert preliminary