Prospects for observing and localizing gravitational-wave transients with Advanced LIGO, Advanced Virgo and KAGRA

Abstract

We present possible observing scenarios for the Advanced LIGO, Advanced Virgo and KAGRA gravitational-wave detectors over the next decade, with the intention of providing information to the astronomy community to facilitate planning for multi-messenger astronomy with gravitational waves. We estimate the sensitivity of the network to transient gravitational-wave signals, and study the capability of the network to determine the sky location of the source. We report our findings for gravitational-wave transients, with particular focus on gravitational-wave signals from the inspiral of binary neutron star systems, which are the most promising targets for multi-messenger astronomy. The ability to localize the sources of the detected signals depends on the geographical distribution of the detectors and their relative sensitivity, and [Formula: see text] credible regions can be as large as thousands of square degrees when only two sensitive detectors are operational. Determining the sky position of a significant fraction of detected signals to areas of 5-[Formula: see text] requires at least three detectors of sensitivity within a factor of [Formula: see text] of each other and with a broad frequency bandwidth. When all detectors, including KAGRA and the third LIGO detector in India, reach design sensitivity, a significant fraction of gravitational-wave signals will be localized to a few square degrees by gravitational-wave observations alone.

Keywords: Data analysis; Electromagnetic counterparts; Gravitational waves; Gravitational-wave detectors.

Fig. 1

Regions of aLIGO (top left), AdV (top right) and KAGRA (bottom) target strain sensitivities as a function of frequency. The binary neutron star (BNS) range, the average distance to which these signals could be detected, is given in megaparsec. Current notions of the progression of sensitivity are given for early, mid and late commissioning phases, as well as the design sensitivity target and the BNS-optimized sensitivity. While both dates and sensitivity curves are subject to change, the overall progression represents our best current estimates

Fig. 2

The planned sensitivity evolution and observing runs of the aLIGO, AdV and KAGRA detectors over the coming years. The colored bars show the observing runs, with the expected sensitivities given by the data in Fig. 1 for future runs, and the achieved sensitivities in O1 and in O2. There is significant uncertainty in the start and end times of planned the observing runs, especially for those further in the future, and these could move forward or backwards relative to what is shown above. The plan is summarised in Sect. 2.2

Fig. 3

Offline transient search results for the first observing run: the cumulative number of triggers with false alarm rates (FARs) smaller than the abscissa value. The dashed line shows the expected noise-only distribution, and the dotted lines show the $90\%$ confidence interval assuming no signals. Potential signals are identified by having smaller FARs than expected. The plots are truncated at a minimum FAR of ${10}^{-2}{\mathrm{yr}}^{-1}$. Left: Compact binary coalescence search results (Abbott et al. 2016c, q). We show results from two search algorithms, GstLAL (Cannon et al. ; Privitera et al. ; Messick et al. 2017) and PyCBC (Dal Canton et al. ; Usman et al. 2016). The most significant triggers for both are LVT151012, GW151226 and GW150914; GW150914 and GW151226 have FARs less than ${10}^{-2}{\mathrm{yr}}^{-1}$. Right: Burst search results (Abbott et al. 2017b). We show results from three search algorithms, coherent Wave Burst (cWB; Klimenko et al. 2016, 2008), Omicron–LALInferenceBurst (oLIB; Lynch et al. 2017) and BayesWave follow-up of cWB (cWB+BW; Kanner et al. 2016). All three found GW150914 (the only cWB trigger above the BayesWave follow-up threshold) with a FAR less than ${10}^{-2}{\mathrm{yr}}^{-1}$. GW151226 and LVT151012 fall below the burst search’s detection threshold

Fig. 4

Source localization by timing triangulation for the aLIGO–AdV network. The locations of the three detectors are indicated by black dots, with LIGO Hanford labeled H, LIGO Livingston as L, and Virgo as V. The locus of constant time delay (with associated timing uncertainty) between two detectors forms an annulus on the sky concentric about the baseline between the two sites (labeled by the two detectors). For three detectors, these annuli may intersect in two locations. One is centered on the true source direction (S), while the other ($S^{\prime }$) is its mirror image with respect to the geometrical plane passing through the three sites. For four or more detectors there is a unique intersection region of all of the annuli. Image adapted from Chatterji et al. (2006)

Fig. 5

Posterior probability density for sky location for an example simulated binary neutron star coalescence observed with a two-detector network. Left: Localization produced by the low-latency bayestar code (Singer et al. ; Singer and Price 2016). Right: Localization produced by the higher-latency (neglecting spin) LALInference (Veitch et al. 2015), which also produces posterior estimates for other parameters. These algorithms are discussed in Sect. 3.2.1, and agreement between them shows that the low-latency localization is comparable to the one produced by the higher-latency pipelines. The star indicates the true source location. The source is at a distance of $266\mathrm{Mpc}$ and has a network signal-to-noise ratio of ${\rho }_c=13.2$ using a noise curve appropriate for the first aLIGO run (O1, see Sect. 4.1). The plot is a Mollweide projection in geographic coordinates. Figure reproduced with permission from Berry et al. (2015), copyright by AAS; further mock sky localizations for the first two observing runs can be found at www.ligo.org/scientists/first2years/ for binary neutron star signals and www.ligo.org/scientists/burst-first2years/ for burst signals

Fig. 6

Anticipated binary neutron star sky localization during the first two observing runs (top: O1, see Sect. 4.1; bottom: O2, see Sect. 4.2). Detector sensitivities were taken to be in the middle of the early (for aLIGO in O1 and AdV in O2) and mid (for aLIGO in O2) bands of Fig. 1. The plots show the cumulative fractions of events with sky-localization areas smaller than the abscissa value. Left: Sky area of $90\%$ credible region ${\mathrm{CR}}_{0.9}^{\mathrm{BNS}}$, the (smallest) area enclosing $90\%$ of the total posterior probability. Right: Searched area $A_{\ast }^{\mathrm{BNS}}$, the area of the smallest credible region containing the true position. Results are shown for the low-latency bayestar (Singer and Price 2016) and higher-latency (neglecting spin) LALInference (LI; Veitch et al. 2015) codes. The O2 results are divided into those where two detectors (2-det) are operating in coincidence, and those where three detectors (3-det) are operating: assuming a duty cycle of 70–$75\%$ for each instrument, the two-detector network would be operating for 42–$44\%$ of the total time and the three-detector network 34–$42\%$ of the time. The shaded areas indicate the $68\%$ confidence intervals on the cumulative distributions. A detection threshold of a signal-to-noise ratio of 12 is used and results are taken from Berry et al. (2015), Singer et al. (2014)

Fig. 7

Simulated sky localization for Gaussian (G; top left), sine–Gaussian (SG; top right), broadband white-noise (WN; bottom left) and binary black hole (BBH; bottom right) bursts during the first two observing runs (O1, see Sect. 4.1, and O2, see Sect. 4.2). The plots show the cumulative fractions of events with searched areas $A_{\ast }$ smaller than the abscissa value. Results are shown for the low-latency coherent WaveBurst (CWB; Klimenko et al. 2005, 2008, 2016), and higher-latency LALInferenceBurst (LIB; Veitch et al. 2015) and BayesWave (Cornish and Littenberg 2015) codes. The O2 results consider only a three-detector (3-det) network; assuming an instrument duty cycle of 70–$75\%$, this would be operational 34–$42\%$ of the time. The BayesWave results are only for O1 and include only events that could be detected by the code. The shaded areas indicate the $68\%$ confidence intervals on the cumulative distributions. A detection threshold of a false alarm rate of approximately $1{\mathrm{yr}}^{-1}$ is used for cWB and LIB, and BayesWave is run as a follow-up for cWB triggers. Results are taken from Essick et al. (2015) and Bécsy et al. (2016)

Fig. 8

Sky localization of a signal with parameters consistent with those for GW150914. The lines enclose the $90\%$ credible regions with different detector networks. Dark blue is for the O1 two-detector network; light blue is for the same Hanford–Livingston network at design sensitivity; red is for the three-detector network including Virgo, with all detectors at early sensitivity, similar to what was expected for O2, and black is for the three detector network at design sensitivity. The plot is an orthographic projection with right ascension measured in hours and declination measured in degrees. Results taken from Gaebel and Veitch (2017)