Towards improving the prospects for coordinated gravitational-wave and electromagnetic observations
aa r X i v : . [ a s t r o - ph . H E ] O c t New Horizons in Time-domain AstronomyProceedings IAU Symposium No. 285, 2011E. Griffin, R. Hanisch, & R. Seaman, eds. c (cid:13) Towards improving the prospects forcoordinated gravitational-wave andelectromagnetic observations
Ilya Mandel , Luke Z. Kelley and Enrico Ramirez-Ruiz School of Physics and Astronomy, University of Birmingham, Edgbaston, B15 2TT, UKemail: [email protected] Harvard-Smithsonian Center for Astrophysics, 60 Garden Street, Cambridge, MA 02138 Department of Astronomy and Astrophysics, University of California, Santa Cruz, CA 95064
Abstract.
We discuss two approaches to searches for gravitational-wave (GW) and electromag-netic (EM) counterparts of binary neutron star mergers. The first approach relies on triggeringarchival searches of GW detector data based on detections of EM transients. We introduce aquantitative approach to evaluate the improvement to GW detector reach due to the extrainformation gained from the EM transient and the increased confidence in the presence of asignal from a binary merger. We also advocate utilizing other transients in addition to shortgamma ray bursts. The second approach involves following up GW candidates with targeted EMobservations. We argue for the use of slower but optimal parameter-estimation techniques to lo-calize the source on the sky, and for a more sophisticated use of astrophysical prior information,including galaxy catalogs, to find preferred followup locations.
Keywords. gravitational waves, stars: neutron, binaries: close
1. Introduction
Advanced ground-based gravitational-wave detectors LIGO [Harry & the LIGO Scientific Collaboration(2010)] and Virgo [Virgo Collaboration (2009)] are expected to begin taking data around2015, operating at a sensitivity approximately a factor of 10 better than their initialcounterparts. Gravitational waves (GWs) emitted during the late inspirals and mergersof compact-object binaries will be one of the main sources for these detectors. Althoughpredictions for merger rates of binary neutron stars and neutron-star – black-hole bi-naries are highly uncertain (see, e.g, Mandel & O’Shaughnessy (2010)), we may expectdetections at a rate between once per few years and a few hundred per year, with perhapsa few tens of detections per year being most likely [Abadie et al. (2010a)].These detections would usher in an era of genuine gravitational-wave astronomy, withGWs being used as another tool to observe the sky. There has been a lot of discussionof the promise of multimessenger observations in the literature (see, e.g., Bloom et al.(2009)). In fact, the recent focus on multimessenger astronomy has been so exclusivethat it is worthwhile to remind ourselves that a significant amount of astrophysics canbe extracted from the GW observations alone, since the GW signal encodes the massesand spins of the binary components, and can be used to probe astrophysics, strong-field gravity, and cosmology even in the absence of electromagnetic (EM) observationsof counterparts to GW events. Nonetheless, there is no doubt that observing both EMand GW counterparts of the same event would be of great astrophysical significance, andwould allow us to settle crucial questions such as the origins of short GRBs.In general, such observations can be achieved in three different ways: (i) through en-tirely independent, serendipitous observations of GW and EM signatures of the same1 Ilya Mandel, Luke Z. Kelley & Enrico Ramirez-Ruizevent, (ii) through triggered searches of archival GW data based on EM transients ob-served during surveys, or (iii) through telescope pointing to search for electromagneticfollowups of GW candidates. Here, we describe some thoughts about future possibilitiesfor triggered GW searches based on observed EM transients, and possible improvementsto recently started efforts to follow up GW triggers with targeted EM observations.
2. GW searches triggered on EM transients
The detection of an electromagnetic transient which may originate from a compact-object binary merger will increase the a priori probability that a given stretch of data fromthe LIGO-Virgo ground-based gravitational-wave detector network contains a signal froma binary coalescence. Additional information contained in the electromagnetic signal, suchas the sky location or distance to the source, can further help to rule out false alarms,and thus lower the necessary threshold for a detection.The LIGO Scientific Collaboration and the Virgo Collaboration have previously usedshort, hard Gamma Ray Bursts (SGRBs) as triggers for searches for compact binarycoalescences in GW detector data [Abadie et al. (2010b)]. SGRBs are believed to be as-sociated with relativistic jets from BNS or NS-BH mergers (see, e.g., Lee & Ramirez-Ruiz(2007)). However, the impact of the SGRB observation on gaining confidence in the pres-ence of a GW signature from a binary merger has not been quantitatively estimated. Infact, most of the observed SGRBs originate too far away to detect the associated GWsignal: the average luminosity distance for the 16 SGRBs with confident host identifi-cations and redshift measurements compiled by Berger (2010) is approximately 5 Gpc,which exceeds the anticipated event horizon of Advanced LIGO by more than a factorof 10. Hence, even when an SGRB is detected, it is unlikely that a gravitational-wavecounterpart is observable, so the detectability threshold is not lowered by as much ascould be anticipated.Kelley et al. (2011) develop a Bayesian framework to accurately incorporate the ad-ditional prior information on the existence of a signal and its timing, sky location, andpotentially distance and inclination, in order to quantify the benefits of triggered searchesin improving the detector reach. The odds ratio for the GW detector data d to containa signal relative to the noise-only model is O ≡ p ( GW | d ) p ( N | d ) = p ( GW ) p ( N ) p ( d | GW ) p ( d | N ) , (2.1)where the first term is the a priori probability of having a GW signal present, and thesecond term is known as the Bayes factor, B = Z p ( θ | GW ) · e
3. EM followups of GW candidate events
The benefits of following up GW candidates with targeted EM observations have longbeen recognized [Finn et al. (1999)]. Recently, the first pilot program for following upGW candidate events has been activated [Abadie et al. (2011)]. There are two significantcomplications in this effort. One is the speed with which GW data can be analyzedto yield information about detection confidence and the sky location of GW candidates.The other is the large positional uncertainty associated with GW detections, which couldencompass tens or even a few hundred square degrees on the sky depending on the SNRof the candidate and the detector network configuration (e.g., Nissanke et al. (2011)).A significant amount of effort has been expended to allow for very rapid processing ofGW data with the aim of achieving latencies of only a few seconds or tens of secondsto identify GW candidates (e.g., Cannon et al. (2011)). Timing triangulation betweendifferent GW interferometers comprising the network has been used so far to rapidlylocalize the source on the sky (e.g., Fairhurst (2009)). Such incoherent use of detector dataallows for rapid analysis, but is likely to yield suboptimal results. The “need for speed”can be over-stated: it is simply not necessary for many of the possible EM counterparts.For example, off-axis optical afterglows outside the jet opening angle will only peak ontime scales greater than ∼ ∼
400 Mpc (the horizon distance of Advanced LIGO for op-timally located and oriented coalescing neutron-star binaries). Finally, some fraction ofmergers may happen outside galaxies altogether, as the progenitor binary could havebeen kicked out of the hosts by supernovae kicks [Kelley et al. (2010)]. An ongoing studyby Vousden et al. (2012) aims to account for these shortcomings by employing moresophisticated astrophysical priors.Despite the improvements mentioned above, finding an electromagnetic counterpart ofa GW candidate will remain extremely challenging, as discussed by Metzger & Berger(2011). Coordinated observing among several facilities may be required to cover the largeuncertainty region with smaller field-of-view instruments. Another alternative that maybe worth investigating is the deployment of a network of inexpensive robotic telescopesspecifically with the goal of following up GW candidates, though it will be difficultto detect any but the closest afterglows. While it is impossible to guarantee that anEM counterpart to given candidate would be found, we can still strive to maximize theprobability of a successful follow up, with a view to ensuring that a sufficient fraction oftriggers are followed up to make at least some multimessenger observations.