Gravitational-wave Merger Forecasting: Scenarios for the early detection and localization of compact-binary mergers with ground based observatories
DDraft version October 20, 2020
Typeset using L A TEX twocolumn style in AASTeX63
Gravitational-wave Merger Forecasting: Scenarios for the early detection and localization ofcompact-binary mergers with ground based observatories
Alexander H. Nitz,
1, 2
Marlin Sch¨afer,
1, 2 and Tito Dal Canton Max-Planck-Institut f¨ur Gravitationsphysik (Albert-Einstein-Institut), D-30167 Hannover, Germany Leibniz Universit¨at Hannover, D-30167 Hannover, Germany Universit´e Paris-Saclay, CNRS/IN2P3, IJCLab, 91405 Orsay, France
ABSTRACTWe present the prospects for the early (pre-merger) detection and localization of compact-binarycoalescences using gravitational waves over the next 10 years. Early warning can enable the directobservation of the prompt and early electromagnetic emission of a neutron star merger. We examinethe capabilities of the ground based detectors at their “Design” sensitivity (2021-2022), the planned“A+” upgrade (2024-2026), and the envisioned “Voyager” concept (late 2020’s). We find that for afiducial rate of binary neutron star mergers of 1000 Gpc − yr − , the Design, A+, and Voyager eranetworks can provide 18, 54, and 195s of warning for one source per year of observing, respectively,with a sky localization area <
100 deg at a 90% credible level. At the same rate, the A+ andVoyager era networks will be able to provide 9 and 43s of warning, respectively, for a source with < localization area. We compare the idealized search sensitivity to that achieved by the PyCBCLive search tuned for pre-merger detection. The gravitational-wave community will be prepared toproduce pre-merger alerts. Our results motivate the operation of observatories with wide fields-of-view,automation, and the capability for fast slewing to observe simultaneously with the gravitational-wavenetwork. Keywords: gravitational waves — neutron stars — compact binary stars INTRODUCTIONThe second generation of gravitational-wave observa-tion began in 2015 with the operation of the twin LIGOobservatories (Aasi et al. 2015). During the first ob-serving run, the first binary black hole mergers were de-tected, which provided insight into gravity in the strong-field regime (Abbott et al. 2016a,b,c). The true eraof gravitational-wave multi-messenger astronomy, how-ever, began with GW170817, the first observation of abinary neutron star (BNS) merger with gravitationalwaves (Abbott et al. 2017a). Only a couple secondsfollowing the gravitational-wave signal observed by theLIGO and Virgo (Acernese et al. 2015) observatories, agamma-ray burst was observed by Fermi-GBM and IN-TEGRAL (Goldstein et al. 2017; Savchenko et al. 2017;Abbott et al. 2017b). About 11 hours later, the opti-cal counterpart was spotted (Coulter et al. 2017) and to
Corresponding author: Alexander H. [email protected] date GW170817 has been observed by over 70 observa-tories spanning the electromagnetic band and includingneutrino and cosmic-ray observatories (see Abbott et al.(2017c) and references therein for a detailed summary).The observation of GW170817 has provided an unprece-dented look into the nuclear equation of state (Abbottet al. 2019a, 2018; Radice et al. 2018; Kiuchi et al. 2019;Capano et al. 2020; Abbott et al. 2020b), the Hub-ble constant (Guidorzi et al. 2017; Hotokezaka et al.2019; Fishbach et al. 2019), the phenomenon of kilo-nova (see Metzger (2020) and references therein), andthe central engine of short gamma-ray bursts (Murguia-Berthier et al. 2020; Wu & MacFadyen 2019; Lazzatiet al. 2020).However, a crucial gap in these observations are therecords of the early time behavior of the optical emis-sion. Optical observations only began hours after theneutron star merger due to the presence of non-Gaussiantransient noise in the LIGO-Livingston data requiringmanual intervention and preventing the initial auto-mated release of a precise sky localization (Abbott et al.2017a). Earlier optical and in particular ultraviolet ob- a r X i v : . [ a s t r o - ph . H E ] O c t servations would have been able to differentiate kilonovaemission models (Arcavi 2018). While the typical la-tency for automated gravitational-wave alerts has beenreduced over time to between tens of seconds to minutesafter merger (Abbott et al. 2019b), the holy grail wouldbe to observe a coalescence’s prompt and early electro-magnetic emission just a matter of seconds after merger.There may be electromagnetic emission which occurs be-fore the merger (Hansen & Lyutikov 2001; Troja et al.2010; Tsang et al. 2011; Metzger & Zivancev 2016; Wanget al. 2016; Wada et al. 2020). A broad summary of thescientific potential of neutron star merger observationscan be found in Burns et al. (2019).To date, the LIGO and Virgo observatories have de-tected dozens of gravitational wave sources (Nitz et al.2019a,c, 2020; Venumadhav et al. 2019a,b; Zackay et al.2019; Abbott et al. 2019c, 2020c,d,e), two BNS merg-ers (Abbott et al. 2017a, 2020a), but only a singlesource, GW170817, had clear electromagnetic counter-parts (Abbott et al. 2017c; Nitz et al. 2019b). During O3there was an active follow-up campaign involving numer-ous telescopes (see e.g. follow-up of GW190425 (LVK2019)), which included but was not limited to follow-up by Swift, ZTF Anand et al. (2020), MASTER, andGRANDMA Antier et al. (2020). However, over thecoming decade, we expect the sensitivity and capabil-ity of ground based gravitational-wave observatories todramatically increase (Abbott et al. 2016d). To matchthis, the infrastructure for both the low-latency (Mes-sick et al. 2017; Adams et al. 2016; Hooper et al. 2012;Klimenko et al. 2016) and pre-merger detection of grav-itational waves is being actively developed by multiplegroups (Cannon et al. 2012; Chu et al. 2016; Kapadiaet al. 2020; Sachdev et al. 2020) with a preliminary testrecently conducted after the end of the third observingrun (O3) (LVK 2020). To take advantage of advancewarning, facilities will need automated operation, wideeffective fields-of-view, and the ability to rapidly point.In this letter, we explore the increasing capability ofthe global gravitational-wave network to detect inspi-ralling binaries seconds to minutes before merger. Weexamine the distribution of detectable sources and theevolution of their distance and sky localization over thenext several years. Finally, we adapt the existing low-latency search PyCBC Live (Nitz et al. 2018a; Dal Can-ton et al. 2020) to gauge if the current search methodswill continue to be suited for pre-merger detection withthe forthcoming global network. PRE-MERGER DETECTION OF MERGERSWITH GRAVITATIONAL WAVES Orbiting compact binaries emit gravitational wavesand, due to the loss of orbital energy, inspiral andeventually merge (Peters 1964). For low mass sources,such as BNSs, this ‘inspiral’ phase of the gravitationalwave signal is the observable portion. The merger andpost-merger gravitational-wave signals are buried in thenoise for current instruments, as they occur at frequen-cies ( ∼ Frequency [Hz] − − − − − A m p li t ud e S p ec t r a l D e n s i t y [ S t r a i n / √ H z ] LIGO-VoyagerLIGO-A+LIGO-DesignAdvVirgoKAGRA
Figure 1.
Noise curves for the Advanced Virgo (red), KA-GRA (purple), and LIGO instruments used in this study.The LIGO instruments are expected to achieve the “Design”curve (green) in the early 2020’s, followed by the “A+” (or-ange) in 2024-2026, and “Voyager” (blue) in the late 2020’s.
Observatories and simulated source population
In this study, we consider the five ground based ob-servatories currently in operation or under construction,namely LIGO-Hanford (H), LIGO-Livingston (L) (Aasiet al. 2015), LIGO-India (I) (Iyer et al. 2011), Virgo(V) (Acernese et al. 2015), and KAGRA (K) (Akutsuet al. 2019). We split our analysis into three sensitiv-ity epochs in different configurations. We denote theseepochs as the “Design” era which covers the expectedsensitivity and operation of the detector network start-ing from 2021-2022, the “A+” era which is timed forthe next planned upgrade to the LIGO instruments ex-pected to begin operation in 2024-2026, with LIGO-India joining with equivalent sensitivity towards towardsthe end of this period (Abbott et al. 2016d), and finallythe “Voyager” era which includes proposed upgrades tothe LIGO instruments predicted to begin operation inthe late 2020’s (LIGO Scientific Collaboration 2017).To assess each detector network, we produce a simu-lated population of O (10 ) BNS mergers, which are uni-formly distributed in volume, and isotropic in binary ori-entation and sky location. For simplicity of comparison,we choose a reference binary with component masses1 . − . (cid:12) . However, as described in Sec. 4, our re-sults can be applied to a more generic population. Thegravitational waveform is calculated using TaylorF2, amodel based on the post-Newtonian approximation toGR (Sathyaprakash & Dhurandhar 1991; Droz et al.1999; Blanchet 2002; Faye et al. 2012) which is suit-able for long duration signals where the merger happensat (cid:38)
500 Hz. Each simulated source is added to Gaus- sian noise colored with the power spectral density cor-responding to each instrument at a particular epoch.There is significant uncertainty in the actual noise ca-pability that will be achieved by each instrument overtime. For the LIGO-Hanford and Livingston detectors,we use the “Design”, “A+”, and “Voyager” noise curvesconsistent with Abbott et al. (2016d); Barsotti et al.(2018); Hall (2019). As done in Abbott et al. (2016d),we assume LIGO-India will join the network in the mid2020s using the “A+” configuration, and from then onwill match the sensitivity of the other LIGO observa-tories. Note, that we use the Virgo design curve in allcases, consistent with the conservative projection fromthe mid 2020’s in Abbott et al. (2016d). For KAGRA,we use its design curve (Kagra 2016). Future upgradesto Virgo and KAGRA in the late 2020’s may increasetheir sensitivity during the Voyager era beyond what weconsider here. A comparison of these noise curves isshown in Fig. 1.The advance warning capabilities of the network willdepend on the instruments meeting their sensitivity tar-gets at low frequencies. For example, if the “Voyager”era instruments only match the sensitivity of the “A+”instruments below a gravitational-wave frequency of 30Hz (a sensitivity reduction of ∼ − . × in this band),then the detection and localization capabilities at timesearlier than ∼
60s before merger will only match thosepredicted for the “A+” era. Closer to merger, as more ofthe signal-to-noise is accumulated from higher frequen-cies, the results would converge to those we show in thenext section, assuming the predicted high frequency sen-sitivity is obtained.2.2.
Source detection and localization
We consider two criteria to define whether a particu-lar simulated source is detected at a given time beforemerger, and hence measure the capabilities of future de-tector networks.The first criterion is an idealized, simplified analy-sis which detects any signal having a total networkSNR >
10. This choice is consistent with the thresh-old for confidently detected mergers in Abbott et al.(2019c); Nitz et al. (2019c). In practice, we may expecta marginally lower threshold (i.e. higher sensitivity), de-pendent on the rate of confounding non-Gaussian noisetransients in future detector networks. For nearly Gaus-sian data, a threshold of ∼ ∼ − . ∼ ∼ ∼ DETECTION AND LOCALIZATIONCAPABILITIESThrough the simulations described in Section 2 we ob-tain the search sensitivity, expected rate of detections,and sky localization capabilities as a function of time be-fore merger. These are shown in Figs. 2, 3, and 4 for the“Design”, “A+”, and “Voyager” era networks, respec-tively. For each era, we compare the reduced “HLV” de-tector network to the full network appropriate for thatera. Note that the times shown in the horizontal axesdo not include the latency of the analysis, which we ex-pect to vary over the years as technical improvementsare made.We find that the PyCBC Live low-latency search is al-ready comparable to the idealized search, though someimprovement may be possible for network configura-tions with a large number of detectors, whereas forthree-detector configurations PyCBC Live already out-performs our simplified analysis when operating at falsealarm rate of 1 per year. We can expect further im-provements in pre-merger analyses to be made, but it isalready clear that existing searches will be fully capableof meeting our predictions throughout the decade, as-suming detector noise quality is comparable to previousobservation runs. S e a r c h R a n g e [ M p c ] HLV
PyCBC LiveIdealized ( ρ c > ) HLVK S k y A r e a [ d e g ] upper limit upper limit upper limit S o u r ce D i s t a n ce [ M p c ] % upper limit90 % upper limit, sky < % upper limit, sky <
100 deg % upper limit, sky <
10 deg − − − − − − −
20 0
Time from Merger [s] − − D e t ec t i o n R a t e [ y r − ] All Detectablesky area < sky area <
100 deg sky area <
10 deg sky area < − − − − − − −
20 0
Time from Merger [s]Credible Level of Sky Area
Figure 2. “Design” era (2021-2022) detection and localization for the HLV network (left) and the full gravitational-wavedetector network (right) as a function of time before merger for a fiducial 1.4-1.4M (cid:12)
BNS merger. (Top) The sky-averageddetection range for the idealized search and PyCBC Live operating at a false alarm rate of once per year. (Middle) The upperlimit on the localization sky area and source distance, respectively, for detectable sources. Sky areas are quoted at the 90%credible level. (Bottom) The detection rate of all sources (black) and those that also have a sky localization less than 1000 deg (blue), 100 deg (orange), 10 deg (green), or 1 deg at a 90% (solid), 50% (dashed), and 25% credible level (dotted). S e a r c h R a n g e [ M p c ] HLV
PyCBC LiveIdealized ( ρ c > ) HLVKI S k y A r e a [ d e g ] upper limit upper limit upper limit S o u r ce D i s t a n ce [ M p c ] % upper limit90 % upper limit, sky < % upper limit, sky <
100 deg % upper limit, sky <
10 deg − − − − −
50 0
Time from Merger [s] − − D e t ec t i o n R a t e [ y r − ] All Detectablesky area < sky area <
100 deg sky area <
10 deg sky area < − − − − −
50 0
Time from Merger [s]Credible Level of Sky Area
Figure 3. “A+” era (2024-2026) detection and localization for the HLV network (left) and the full gravitational-wave detectornetwork (right) as a function of time before merger for a fiducial 1.4-1.4M (cid:12)
BNS merger. (Top) The sky-averaged detectionrange for the idealized search and PyCBC Live operating at a false alarm rate of once per year. (Middle) The upper limit onthe localization sky area and source distance, respectively, for detectable sources. Sky areas are quoted at the 90% crediblelevel. (Bottom) The detection rate of all sources (black) and those that also have a sky localization less than 1000 deg (blue),100 deg (orange), 10 deg (green), or 1 deg at a 90% (solid), 50% (dashed), and 25% credible level (dotted). S e a r c h R a n g e [ M p c ] HLV
PyCBC LiveIdealized ( ρ c > ) HLVKI S k y A r e a [ d e g ] upper limit upper limit upper limit S o u r ce D i s t a n ce [ M p c ] % upper limit90 % upper limit, sky < % upper limit, sky <
100 deg % upper limit, sky <
10 deg − − − − − −
100 0
Time from Merger [s] − − D e t ec t i o n R a t e [ y r − ] All Detectablesky area < sky area <
100 deg sky area <
10 deg sky area < − − − − − −
100 0
Time from Merger [s]Credible Level of Sky Area
Figure 4. “Voyager” era (late 2020’s) detection and localization for the HLV network (left) and the full gravitational-wavedetector network (right) as a function of time before merger for a fiducial 1.4-1.4M (cid:12)
BNS merger. (Top) The sky-averageddetection range for the idealized search and PyCBC Live operating at a false alarm rate of once per year. (Middle) The upperlimit on the localization sky area and source distance, respectively, for detectable sources. Sky areas are quoted at the 90%credible level. (Bottom) The detection rate of all sources (black) and those that also have a sky localization less than 1000 deg (blue), 100 deg (orange), 10 deg (green), or 1 deg at a 90% (solid), 50% (dashed), and 25% credible level (dotted). For 1 . − . (cid:12) BNS mergers and a merger rate den-sity of ∼ − yr − (Abbott et al. 2019c, 2020a),we expect to detect one source per year with a 90% cred-ible sky localization area <
100 deg at 18, 54, or 195 sbefore merger for the “Design”, “A+”, and “Voyager”full networks, respectively. We note that detailed studieswill be needed to determine the optimal follow-up strat-egy for specific observatories, however, we can exploresome of the generic choices. Assuming a fixed observa-tion sky area, facilities that would be willing to observea higher rate of sources, and accept that the true sourcelocation may be outside the observed region a fractionof the time, will observe significantly more counterparts.For example, if an observatory targets every 50% credi-ble region with area less than 100 deg the warning timeis either increased to 34, 104, and 335 s, respectively, oralternatively at the same warning times discussed ear-lier, we can expect ∼ − <
10 deg . For the 50% credible region, we can increasethe candidate rate to ∼ − <
100 deg with 17 (18), 54 (54),and 178 (194) s of warning, respectively. There is littledifference between the 100 and 200 Mpc cases as thevast majority of well localized sources will be at closedistances. At 60s before merger, 90% of the detectedsources will be closer than 35 (13), 81 (28), and 231 (71)Mpc for the “Design”, “A+”, and “Voyager” networks,respectively, if we require that the source be localizedwith a 90% credible area <
100 (10) deg .For systems detected at the earliest possible times con-sidered, the sky localization typically evolves from anearly multimodal distribution to a final unimodal or bi-modal distribution, which is orders of magnitude moreprecise than the initial one. It is not uncommon forthe initial localization to also be unimodal, though, butstill orders of magnitude less precise than the final one.However, if we select cases where the initial localizationis more precise than ∼
100 deg , we find most cases arealready unimodal at the earliest time, i.e. they have aconsistent overall direction throughout the inspiral. APPLICATION TO OTHER SOURCESWhile for simplicity, we have reported results for afiducial 1 . − . (cid:12) BNS merger, our results can bestraightforwardly applied to other sources by scaling ofthe time and sensitive distance (or rate/volume as ap-propriate). The time axis scales inversely with the totalmass of the source so that T m ,m = T . − . . (cid:12) m + m (1)where T . − . is a time from our figures and m , arethe desired source’s redshifted component masses. Thistime rescaling directly accounts for the difference in lo-calization for different mass sources, as the sky localiza-tion area is only dependent on the frequency bandwidthand the detector configuration for long duration signals,where the merger is above the detectors’ sensitive fre-quency band.The signal amplitude scales as the th power of thesource’s chirp mass, which implies that volume and de-tection rate scale as R m ,m = R . − . ( m m ) / ( m + m ) / . / . (2)where R . − . is the rate of detections shown in our fig-ures at a merger rate of 1000 Gpc − yr − . For illustra-tive purposes, if we assume that the rate of 1 . − . (cid:12) sources were 100 Gpc − yr − (which is consistent withlimits reported in Abbott et al. (2019c)), then we’d ex-pect for the Voyager era to be able to have 70 secondsof warning for about one source per year with sky area <
100 deg .This same scaling may also be applied for heavy bi-nary black hole mergers, as long as we only considertimes before merger. After this time, the sky localiza-tion distribution is no longer accounted for by a simpletime rescaling due to the signal terminating within themost sensitive frequency band. For instance, consider-ing GW190521 (Abbott et al. 2020f), which may havemerged within the accretion disk of a supermassive blackhole and produced an optical counterpart (Graham et al.2020), we find that the time scale factor is ∼
50. Hence,even in an optimistic Voyager era, we could expect nomore than a few seconds warning for similar mergers. CONCLUSIONSAchieving the goal of the prompt electromagnetic ob-servation of a compact binary merger requires coordina-tion across different observatories and cutting-edge in-struments and facilities with wide fields of view, rapidpointing, and fully automated operation. By simulatinga population of neutron star mergers, and the analy-sis of the associated data with current technology, wehave shown that over the next decade the pre-mergerwarning time may increase by an order of magnitudefrom O(10) to O(100) seconds. For many telescopes,this will not yet be sufficient to re-point and tile a 100deg area (Coughlin et al. 2019a), although notable ex-ceptions exist (Gehrels et al. 2004; Sagiv et al. 2014),including Swift (Tohuvavohu et al. 2020), ZTF (Bellmet al. 2018; Coughlin et al. 2019b), MASTER (Kornilovet al. 2012), and the CTA (Acharya et al. 2013). Variousfacilities may also be able to use pre-merger warnings toalter triggering or observing configurations (James et al.2019).It is our hope that with the roadmap we provide, theobserving community can plan for continued and auto-mated operation of existing observatories, and envisionbold new missions with varied observation bands andthe goal of the first forecasted observation of a BNSmerger within this decade. This includes concepts suchas the Transient Astrophysics Probe (Camp & TAPTeam 2019). As GW170817 introduced gravitationalwaves to the field of multimessenger astronomy, we ex- pect a multimessenger, multiband, prompt observationof a neutron star merger to be an important milestonein rapid time domain astronomy.Data associated with the simulations is releasedat https://github.com/gwastro/gw-merger-forecasting.ACKNOWLEDGMENTSWe thank Aaron Tohuvavohu, Eric Burns, MichaelCoughlin and Nelson Christensen for their comments.This work was spurred by discussions and ideas at theAspen Center for Physics, which is supported by Na-tional Science Foundation grant PHY-1607611. Weacknowledge the Max Planck Gesellschaft and the Atlascluster computing team at AEI Hannover for support.Software: Numpy (van der Walt et al. 2011), Scipy(Virtanen et al. 2020), Astropy (Robitaille et al. 2013),Matplotlib (Hunter 2007), ligo.skymap (Singer 2018),PyCBC (Nitz et al. 2018b).REFERENCES Aasi, J., et al. 2015, Class. Quantum Grav., 32, 074001,doi: 10.1088/0264-9381/32/7/074001Abbott, B., et al. 2017a, Phys. Rev. Lett., 119, 161101,doi: 10.1103/PhysRevLett.119.161101—. 2017b, Astrophys. J. Lett., 848, L13,doi: 10.3847/2041-8213/aa920c—. 2018, Phys. Rev. Lett., 121, 161101,doi: 10.1103/PhysRevLett.121.161101—. 2020a, Astrophys. J. Lett., 892, L3,doi: 10.3847/2041-8213/ab75f5Abbott, B. P., et al. 2016a, Phys. Rev. Lett., 116, 061102,doi: 10.1103/PhysRevLett.116.061102—. 2016b, Phys. Rev. Lett., 116, 221101,doi: 10.1103/PhysRevLett.116.221101—. 2016c, Phys. Rev. X, 6, 041015, doi: 10.1103/PhysRevX.6.041015,10.1103/PhysRevX.8.039903—. 2016d, Living Rev. Relat., 19, 1, doi: 10.1007/lrr-2016-1—. 2017c, Astrophys. J., 848, L12,doi: 10.3847/2041-8213/aa91c9—. 2019a, Phys. Rev. X, 9, 011001,doi: 10.1103/PhysRevX.9.011001—. 2019b, Astrophys. J., 875, 161,doi: 10.3847/1538-4357/ab0e8f—. 2019c, Phys. Rev. X, 9, 031040,doi: 10.1103/PhysRevX.9.031040—. 2020b, Classical and Quantum Gravity, 37, 045006,doi: 10.1088/1361-6382/ab5f7c Abbott, R., Abbott, T. D., Abraham, S., et al. 2020c, Phys.Rev. D, 102, 043015, doi: 10.1103/PhysRevD.102.043015Abbott, R., et al. 2020d, The Astrophysical Journal, 896,L44, doi: 10.3847/2041-8213/ab960f—. 2020e, Phys. Rev. Lett., 125, 101102,doi: 10.1103/PhysRevLett.125.101102—. 2020f, Phys. Rev. Lett., 125, 101102,doi: 10.1103/PhysRevLett.125.101102Acernese, F., et al. 2015, Class. Quantum Grav., 32,024001, doi: 10.1088/0264-9381/32/2/024001Acharya, B., et al. 2013, Astropart. Phys., 43, 3,doi: 10.1016/j.astropartphys.2013.01.007Adams, T., Buskulic, D., Germain, V., et al. 2016, Class.Quant. Grav., 33, 175012,doi: 10.1088/0264-9381/33/17/175012Akutsu, T., et al. 2019, Nature Astron., 3, 35,doi: 10.1038/s41550-018-0658-yAnand, S., et al. 2020, doi: 10.1038/s41550-020-1183-3Antier, S., et al. 2020, Mon. Not. Roy. Astron. Soc., 497,5518, doi: 10.1093/mnras/staa1846Arcavi, I. 2018, Astrophys. J. Lett., 855, L23,doi: 10.3847/2041-8213/aab267Barsotti, L., McCuller, L., Evans, M., & Fritchel, P. 2018,The A+ design curve, https://dcc.ligo.org/public/0149/T1800042/005/T1800042-v5.pdf Bellm, E. C., Kulkarni, S. R., Graham, M. J., et al. 2018,Publications of the Astronomical Society of the Pacific,131, 018002, doi: 10.1088/1538-3873/aaecbeBlanchet, L. 2002, Living Rev. Rel., 5, 3—. 2014, Living Rev. Rel., 17, 2Blanchet, L., & Damour, T. 1989, Ann. Inst. H. PoincarePhys. Theor., 50, 377Brown, D. A. 2004, PhD thesis, University ofWisconsin–MilwaukeeBrown, D. A., Harry, I., Lundgren, A., & Nitz, A. H. 2012,Phys.Rev., D86, 084017,doi: 10.1103/PhysRevD.86.084017Burns, E., et al. 2019. https://arxiv.org/abs/1903.03582Camp, J., & TAP Team. 2019, in The Space AstrophysicsLandscape for the 2020s and Beyond, Vol. 2135, 5027Cannon, K., Cariou, R., Chapman, A., et al. 2012,Astrophys.J., 748, 136,doi: 10.1088/0004-637X/748/2/136Capano, C. D., Tews, I., Brown, S. M., et al. 2020, NatureAstron., 4, 625, doi: 10.1038/s41550-020-1014-6Chu, Q., Howell, E., Rowlinson, A., et al. 2016, Mon. Not.Roy. Astron. Soc., 459, 121, doi: 10.1093/mnras/stw576Clark, J. A., Bauswein, A., Stergioulas, N., & Shoemaker,D. 2016, Class. Quant. Grav., 33, 085003,doi: 10.1088/0264-9381/33/8/085003Coughlin, M. W., et al. 2019a, Mon. Not. Roy. Astron.Soc., 489, 5775, doi: 10.1093/mnras/stz2485—. 2019b, Astrophys. J. Lett., 885, L19,doi: 10.3847/2041-8213/ab4ad8Coulter, D. A., et al. 2017, Science,doi: 10.1126/science.aap9811Dal Canton, T., Nitz, A. H., Gadre, B., et al. 2020.https://arxiv.org/abs/2008.07494Droz, S., Knapp, D. J., Poisson, E., & Owen, B. J. 1999,PhRvD, 59, 124016Faye, G., Marsat, S., Blanchet, L., & Iyer, B. R. 2012,Class. Quant. Grav., 29, 175004,doi: 10.1088/0264-9381/29/17/175004Fishbach, M., et al. 2019, Astrophys. J. Lett., 871, L13,doi: 10.3847/2041-8213/aaf96eGehrels, N., et al. 2004, Astrophys. J., 611, 1005,doi: 10.1086/422091Goldstein, A., et al. 2017, Astrophys. J., 848, L14,doi: 10.3847/2041-8213/aa8f41Graham, M., et al. 2020, Phys. Rev. Lett., 124, 251102,doi: 10.1103/PhysRevLett.124.251102Guidorzi, C., et al. 2017, Astrophys. J. Lett., 851, L36,doi: 10.3847/2041-8213/aaa009 Hall, E. 2019, Horizon plots and noise curves for second-and third-generation detectors,https://dcc.ligo.org/LIGO-T1800084-v5/publicHansen, B. M., & Lyutikov, M. 2001, Mon. Not. Roy.Astron. Soc., 322, 695,doi: 10.1046/j.1365-8711.2001.04103.xHooper, S., Chung, S. K., Luan, J., et al. 2012, Phys. Rev.D, 86, 024012, doi: 10.1103/PhysRevD.86.024012Hotokezaka, K., Nakar, E., Gottlieb, O., et al. 2019, NatureAstron., 3, 940, doi: 10.1038/s41550-019-0820-1Hunter, J. D. 2007, Computing in Science & Engineering, 9,90, doi: 10.1109/MCSE.2007.55Iyer, B., et al. 2011, LIGO India,https://dcc.ligo.org/LIGO-M1100296/publicJames, C. W., Anderson, G. E., Wen, L., et al. 2019, Mon.Not. Roy. Astron. Soc., 489, L75,doi: 10.1093/mnrasl/slz129Kagra. 2016, Kagra Design Sensitivity Noise Curve, https://github.com/lscsoft/lalsuite/blob/master/lalsimulation/lib/LIGO-T1600593-v1-KAGRA Design.txtKapadia, S. J., Singh, M. K., Shaikh, M. A., Chatterjee, D.,& Ajith, P. 2020, Astrophys. J. Lett., 898, L39,doi: 10.3847/2041-8213/aba42dKiuchi, K., Kyutoku, K., Shibata, M., & Taniguchi, K.2019, The Astrophysical Journal, 876, L31,doi: 10.3847/2041-8213/ab1e45Klimenko, S., et al. 2016, Phys. Rev. D, 93, 042004,doi: 10.1103/PhysRevD.93.042004Kornilov, V. G., Lipunov, V. M., Gorbovskoy, E. S., et al.2012, Experimental Astronomy, 33, 173,doi: 10.1007/s10686-011-9280-zLazzati, D., Ciolfi, R., & Perna, R. 2020, Astrophys. J.,898, 59, doi: 10.3847/1538-4357/ab9a44LIGO Scientific Collaboration. 2017, Instrument ScienceWhite Paper, https://dcc.ligo.org/public/0142/T1700231/003/T1700231-v3.pdfLVK. 2019, GCN, 24185, 1.https://gcn.gsfc.nasa.gov/other/GW190425z.gcn3LVK. 2020, Test of early warning alerts, https://emfollow.docs.ligo.org/userguide/early warning.htmlMessick, C., Blackburn, K., Brady, P., et al. 2016.https://arxiv.org/abs/1604.04324Messick, C., et al. 2017, Phys. Rev. D, 95, 042001,doi: 10.1103/PhysRevD.95.042001Metzger, B. D. 2020, Living Rev. Rel., 23, 1,doi: 10.1007/s41114-019-0024-0Metzger, B. D., & Zivancev, C. 2016, Mon. Not. Roy.Astron. Soc., 461, 4435, doi: 10.1093/mnras/stw1800Murguia-Berthier, A., Ramirez-Ruiz, E., De Colle, F., et al.2020. https://arxiv.org/abs/2007.122451