Multi-Messenger Astrophysics

TL;DR

Multi-messenger astrophysics combines photons, gravitational waves, neutrinos, and cosmic rays to gain comprehensive insights into cosmic sources, with recent breakthroughs including the detection of gravitational waves and high-energy neutrinos.

Contribution

This review summarizes recent achievements in multi-messenger astrophysics and discusses future directions and the development of next-generation observatories.

Findings

Detection of ultra-high energy cosmic rays spectrum.

Discovery of diffuse high energy neutrino background.

First direct detections of gravitational waves.

Abstract

Multi-messenger astrophysics, a long-anticipated extension to traditional and multiwavelength astronomy, has recently emerged as a distinct discipline providing unique and valuable insights into the properties and processes of the physical universe. These insights arise from the inherently complementary information carried by photons, gravitational waves, neutrinos, and cosmic rays about individual cosmic sources and source populations. Realizing the observation of astrophysical sources via non-photonic messengers has presented enormous challenges, as evidenced by the fiscal and physical scales of the multi-messenger observatories. However, the scientific payoff has already been substantial, with even greater rewards promised in the years ahead. In this review we survey the current status of multi-messenger astrophysics, highlighting some exciting recent results, and addressing the major follow-on questions they have raised. Key recent achievements include the measurement of the spectrum of ultra-high energy cosmic rays out to the highest observable energies; discovery of the diffuse high energy neutrino background; the first direct detections of gravitational waves and the use of gravitational waves to characterize merging black holes and neutron stars in strong-field gravity; and the identification of the first joint electromagnetic + gravitational wave and electromagnetic + high-energy neutrino multi-messenger sources. We then review the rationales for the next generation of multi-messenger observatories, and outline a vision of the most likely future directions for this exciting and rapidly advancing field.

Equations

$$p_{cr}+p_t/{\gamma }_t\to p/n+{\pi }^{\pm }+{\pi }^0+K^{\pm }+\cdots$$

$${\pi }^{+}\to$$

$${\mu }^{+}\to e^{+}+{\nu }_e+{\bar{\nu }}_{\mu }$$

$${\pi }^{-}\to$$

$${\mu }^{-}\to e^{-}+{\bar{\nu }}_e+{\nu }_{\mu }$$

$${\pi }^0\to$$

$$K^{+}/K^{-}\to$$

$$n\to$$

Full text

**Multi-Messenger Astrophysics111Update of arXiv:1906.10212, v1, which appeared in Nature Reviews Physics, 1:585 (2019), https://www.nature.com/articles/s42254-019-0101-z **

Péter Mészáros1,2,3,4, Derek B. Fox1,3,4, Chad Hanna2,1,3,4, Kohta Murase2,1,3,4,5

1Dept. of Astronomy & Astrophysics, Pennsylvania State University, University Park, PA 16802, USA

2Dept. of Physics, Pennsylvania State University, University Park, PA 16802, USA

3Center for Particle and Gravitational Astrophysics, Pennsylvania State University, University Park, PA 16802, USA

4Institute for Gravitation and the Cosmos, Pennsylvania State University, University Park, PA 16802, USA

5Center for Gravitational Physics, Yukawa Institute for Theoretical Physics, Kyoto University, Kyoto 606-8502 Japan

**Abstract

**Multi-messenger astrophysics, a long-anticipated extension to traditional and multiwavelength astronomy, has recently emerged as a distinct discipline providing unique and valuable insights into the properties and processes of the physical universe. These insights arise from the inherently complementary information carried by photons, gravitational waves, neutrinos, and cosmic rays about individual cosmic sources and source populations. This complementarity is the basic reason why multi-messenger astrophysics is much more than just the sum of the parts. Realizing the observation of astrophysical sources via non-photonic messengers has presented enormous challenges, as evidenced by the fiscal and physical scales of the multi-messenger observatories. However, the scientific payoff has already been substantial, with even greater rewards promised in the years ahead. In this review we survey the current status of multi-messenger astrophysics, highlighting some exciting recent results, and addressing the major follow-on questions they have raised. Key recent achievements include the measurement of the spectrum of ultra-high energy cosmic rays out to the highest observable energies; discovery of the diffuse high energy neutrino background; the first direct detections of gravitational waves and the use of gravitational waves to characterize merging black holes and neutron stars in strong-field gravity; and the identification of the first joint electromagnetic + gravitational wave and electromagnetic + high-energy neutrino multi-messenger sources. We then review the rationales for the next generation of multi-messenger observatories, and outline a vision of the most likely future directions for this exciting and rapidly advancing field.

Key Points

Multi-messenger astrophysics aspires to make use of the information provided about the astrophysical universe by all four fundamental forces of Nature, namely the gravitational, the weak and the strong forces, besides the electromagnetic force which previously had provided almost all our information about the Cosmos. These new channels provide previously untapped, qualitatively different and complementary types of information, capable of probing down to the densest and energy-richest regions of cosmic objects, which were hitherto hidden from astronomers’ sights. 2. 2.

Diffuse backgrounds of high-energy neutrinos (HENs) with energies from \mathrel{\mathchoice{\lower 0.5pt\vbox{\halign{\mathnew\displaystyle\hfil#\hfil\cr>\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\textstyle\hfil#\hfil\cr>\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\scriptstyle\hfil#\hfil\cr>\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\scriptscriptstyle\hfil#\hfil\cr>\crcr\sim\crcr}}}}10 TeV to PeV, ultra-high energy cosmic rays (UHECRs) at energies of \mathrel{\mathchoice{\lower 0.5pt\vbox{\halign{\mathnew\displaystyle\hfil#\hfil\cr>\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\textstyle\hfil#\hfil\cr>\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\scriptstyle\hfil#\hfil\cr>\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\scriptscriptstyle\hfil#\hfil\cr>\crcr\sim\crcr}}}}10^{18}\hbox{~{}eV}, and $\gamma$-rays with energies between MeV and $\sim$TeV have been measured, or upper limits have been provided, by Cherenkov detectors, satellites and ground-based air-shower arrays. 3. 3.

Gravitational waves (GWs) from merging stellar mass black hole and neutron star binaries have been detected at frequencies in the \mathrel{\mathchoice{\lower 0.5pt\vbox{\halign{\mathnew\displaystyle\hfil#\hfil\cr>\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\textstyle\hfil#\hfil\cr>\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\scriptstyle\hfil#\hfil\cr>\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\scriptscriptstyle\hfil#\hfil\cr>\crcr\sim\crcr}}}} 10 Hz to $\sim$ 1 kHz range with laser interferometric gravitational-wave detectors. 4. 4.

The sources of the diffuse UHECR and HEN backgrounds remain unknown, although a gamma-flaring blazar (a type of active galaxy with a massive black hole at the center ejecting a relativistic plasma jet towards the observer) has been tentatively identified with observed HENs. While up to $\sim$85% of the $\gamma$-ray background can be attributed to blazars, it appears that at most 30% of the HEN background can be due to blazars. 5. 5.

Formation channels for the observed stellar mass black hole binaries, and their possible role as a cosmologically relevant component of the dark matter, is currently under debate. 6. 6.

There is a natural physical connection between high energy cosmic ray interactions and the resulting very high energy neutrinos and gamma-rays, which needs to be fully exploited to better understand the nature of their unknown astrophysical sources. The connection with gravitational wave emission, while less direct, can be expected to provide important information about supermassive black hole populations and dynamics. 7. 7.

Even before the arrival of the next generation of gravitational wave, neutrino, and cosmic ray detectors, the present advanced LIGO/VIRGO detectors will be able to detect hundreds of binary mergers up to $\sim$Gpc distances; yet electromagnetic (EM) counterpart searches rely primarily on the aging space-based facilities Swift and Fermi, currently operating well beyond their design lifetimes. These EM counterparts have been found mainly in gamma- or X-rays, and there is an urgent need for a new generation of EM detectors, also extending into other frequencies including the UV, optical/IR, and radio.

1 Introduction

Of the four fundamental forces in nature – the electromagnetic, gravitational, weak and strong nuclear forces – until the middle of the 20th century it was only messengers of the electromagnetic force, in the form of optical photons, which allowed astronomers to study the distant universe. Subsequently, advancing technology added to these radio, infrared, ultra-violet, X-ray and gamma-ray photons. Finally, in the last few decades the messengers of the other three forces, namely gravitational waves (GWs), neutrinos, and cosmic rays (CRs), began to be used in earnest. Thus, we are now finally using the complete set (as far as known) of forces of Nature, which are revealing exciting and hitherto unknown details about the Cosmos and its denizens.

These new non-photonic messengers are generally more challenging to detect and to trace back to their cosmic sources than most electromagnetic emissions. When detected, they are usually associated with extremely high mass or high energy density configurations, e.g. the dense core of normal stars, stellar explosions occurring at the end of the nuclear burning life of massive stars, the surface neighborhood of extremely compact stellar remnants such as white dwarfs, neutron stars or black holes, the strong and fast varying gravitational field near black holes of either stellar mass or the much more massive ones in the core of galaxies, or in energetic shocks in high velocity plasmas associated with such compact astrophysical sources. This association with the most violent astrophysical phenomena known means that the interpretation of multi-messenger observations requires, and can have implications for, our theories of fundamental physics, including strong-field gravity, nuclear physics, and particle interactions.

The study of such high energy compact objects started in earnest in the 1950’s, after decades of a slower build-up with increasingly large ground-based optical telescopes. The first major breakthroughs came from the deployment of large radio-telescopes, followed by the launching of satellites equipped with X-ray and later gamma-ray detectors, which established the existence of active galactic nuclei, neutron stars, and black holes, and revealed dramatic high-energy transient phenomena including X-ray novae, X-ray bursts, and gamma-ray bursts (GRBs). Starting in the late 1960’s, large underground neutrino detectors were built, measuring first the neutrinos produced in the Sun and later those arising from a supernova explosion; and it was only in the current decade that extragalactic neutrinos in the TeV-PeV range were discovered. Cosmic rays in the GeV energy range started being measured in the 1910s, but it was only in the 1960s that large detectors started measuring higher energies implying an extragalactic origin, and only in the last decade has it become practical to start investigating the spectrum and composition in the $10^{18}-10^{20}$ eV range. Gravitational wave detectors started being built in the 1970s, but it was not until the 1990s that new technologies and large enough arrays began to be built approaching the sensitivity required for detections, the first successes starting in 2015. For the first time, detectors covering all four fundamental forces of Nature (Fig, 1) have been thrown into the breach to explore all the previously hidden aspects of the Cosmos.

2 Mono- and Multi-Messenger Advances

The exciting experimental results listed in this section confirm many of the theoretical/phenomenological predictions and expectations that had been formulated over the past several decades, while also bringing up new surprises, which we discuss in the subsequent section.

2.1 Recent Non-Photonic Mono-Messenger Results

$\bullet$  Ultra-High Energy Cosmic Rays.- The Pierre Auger cosmic ray observatory (PAO) [1], located in Argentina, is a 3,000 km2 array of 1660 water Cherenkov stations and 27 air-fluorescence telescopes (one of the tanks and a set of fluorescence telescopes is in Fig. 1, lower right), designed to detect UHECRs at energies between $10^{17}\hbox{~{}eV}$ to $10^{21}\hbox{~{}eV}$. Its measurements of the diffuse UHECR flux energy spectrum, starting in 2009, confirmed conclusively the existence of a spectral steepening setting in near $6\times 10^{19}\hbox{~{}eV}$, e.g. [2], which had been first observed by the HiRes instrument [3]. This is consistent with the so-called GZK (Greisen, Zatsepin, Kuz’min, [4, 5]) feature expected from CR proton energy losses or from heavy ion photo-dissociation [6] due to interactions with cosmic microwave background photons, although it could also be due to reaching a maximum acceleration at the sources. From 2010 onwards, Auger also started showing evidence for an UHECR chemical composition becoming heavier above \mathrel{\mathchoice{\lower 0.5pt\vbox{\halign{\mathnew\displaystyle\hfil#\hfil\cr>\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\textstyle\hfil#\hfil\cr>\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\scriptstyle\hfil#\hfil\cr>\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\scriptscriptstyle\hfil#\hfil\cr>\crcr\sim\crcr}}}}10^{18.5}\hbox{~{}eV}. The statistical significance of these results has become stronger over the years [7, 8, 9]. The spectral results are consistent, within statistical uncertainties with those obtained with the smaller Telescope Array (TA) UHECR observatory [10], which is important because Auger is in the Southern hemisphere while TA is in the Northern. The chemical composition issue [11] is still under debate, although a joint Auger-TA paper [12] shows results which agree within the errors. The angular resolution in the arrival direction of UHECRs is below $1^{o}$ above $\sim 10^{19}\hbox{~{}eV}$ for both protons and heavy elements, although the magnetic deflection increases with mass; around $10^{19}\hbox{~{}eV}$ it is \mathrel{\mathchoice{\lower 0.5pt\vbox{\halign{\mathnew\displaystyle\hfil#\hfil\cr<\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\textstyle\hfil#\hfil\cr<\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\scriptstyle\hfil#\hfil\cr<\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\scriptscriptstyle\hfil#\hfil\cr<\crcr\sim\crcr}}}}5^{o} for protons, while for heavy nuclei it could be tens of degrees. At these energies, due to the energy losses caused by the GZK effect mentioned, these UHECRs must have originated within distances of \mathrel{\mathchoice{\lower 0.5pt\vbox{\halign{\mathnew\displaystyle\hfil#\hfil\cr<\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\textstyle\hfil#\hfil\cr<\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\scriptstyle\hfil#\hfil\cr<\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\scriptscriptstyle\hfil#\hfil\cr<\crcr\sim\crcr}}}}100{\rm Mpc}. So far, all attempts at finding angular spatial correlations between UHECRs and any type of known cosmic sources have been unsuccessful [2].

$\bullet$  High Energy Neutrinos.- The IceCube neutrino observatory [13] consists of a cubic kilometer (roughly a Gigaton) of ice at a depth between 1.4 and 2.4 km below the South Pole, instrumented with 86 strings connecting 5,160 optical phototubes (see schematic in Fig. 1, lower right), which measure the light radiated from charged particles produced by passing high energy neutrinos interacting with the ice. Its construction was finished in 2010, and in 2012-2013 it discovered a diffuse flux of neutrinos in the range 100\hbox{~{}TeV}\mathrel{\mathchoice{\lower 0.5pt\vbox{\halign{\mathnew\displaystyle\hfil#\hfil\cr<\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\textstyle\hfil#\hfil\cr<\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\scriptstyle\hfil#\hfil\cr<\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\scriptscriptstyle\hfil#\hfil\cr<\crcr\sim\crcr}}}}E_{\nu}\mathrel{\mathchoice{\lower 0.5pt\vbox{\halign{\mathnew\displaystyle\hfil#\hfil\cr<\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\textstyle\hfil#\hfil\cr<\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\scriptstyle\hfil#\hfil\cr<\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\scriptscriptstyle\hfil#\hfil\cr<\crcr\sim\crcr}}}}`1\hbox{PeV} [14, 15], later extended down to \mathrel{\mathchoice{\lower 0.5pt\vbox{\halign{\mathnew\displaystyle\hfil#\hfil\cr<\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\textstyle\hfil#\hfil\cr<\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\scriptstyle\hfil#\hfil\cr<\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\scriptscriptstyle\hfil#\hfil\cr<\crcr\sim\crcr}}}}100\hbox{~{}TeV}. The energy spectrum $dN/dE_{\nu}$ can be fitted with a $\sim-2.5$ index power law, but there may be an indication for two components, steeper below $\sim 200\hbox{~{}TeV}$ and flatter (index $\sim-2$ above that, the highest energy so far being $\sim 10\hbox{PeV}$. IceCube detects all neutrino flavors, with muon neutrino charged current interactions resulting in elongated Cherenkov tracks and all other neutrino flavors and interactions largely producing near-spherical optical Cherenkov signals from secondary particle cascades, the direction of arrival being uncertain by $\sim 10^{o}-15^{o}$ for cascades and $\sim 0.5^{o}-1.0^{o}$ for tracks. Tau neutrinos can also be identified at sufficiently higher energies, where the statistics are lower, and these have not yet been identified, although a suggested tau-like candidate has been discussed [16]. The observed flavor distribution is compatible with complete flavor mixing having occurred due to the neutrino oscillation phenomenon over cosmological distances [17, 18]. So far there is no evident correlation of the observed neutrinos with any type of known cosmic objects [19], except for one interesting case discussed below. The smaller underwater Cherenkov telescopes ANTARES [20] and Baikal-GVD [21] have also been in operation and providing upper limits. At much higher energies, the high altitude balloon experiment ANITA [22], flying in a circumpolar orbits in Antarctica, has used a radio technique to measure neutrinos at \mathrel{\mathchoice{\lower 0.5pt\vbox{\halign{\mathnew\displaystyle\hfil#\hfil\cr>\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\textstyle\hfil#\hfil\cr>\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\scriptstyle\hfil#\hfil\cr>\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\scriptscriptstyle\hfil#\hfil\cr>\crcr\sim\crcr}}}}10^{17} eV, which is starting to provide constraints on cosmological neutrino sources and the GZK-related cosmogenic neutrino fluxes, complementary with those provided by Auger [23].

$\bullet$  *Gravitational Waves.- * The Laser Interferometric Gravitational Wave Observatory (LIGO) consist of two detectors, in Louisiana and Washington state, each with 4 km long L-shaped arms, (Fig. 1, upper left) which in 2015 began operation in the $\sim 10-10^{3}$ Hz frequency range [24]. Another array, VIRGO [25], located near Pisa, Italy, and similarly L-shaped with 3 Km long arms, has been operating at epochs coincident with LIGO. Both are actively being commissioned and will achieve design sensitivity in the coming years. The long-awaited first discovery of gravitational waves from a stellar mass binary black hole merger (labeled GW150914) was announced by LIGO in 2016 [26] (Fig. 2(a). This was soon followed by a number of other binary black hole (BBH) mergers detected both by LIGO and, with lower statistical significance, by VIRGO as well [27]. These BHs weigh up to several tens of solar masses, and have low spins. However, despite intensive searches, no other messengers associated with BBH mergers have been detected so far, except for a possible $\gamma$-ray burst [28] in GW150914.

$\bullet$  *Electromagnetic Detections.- * Except for the binary black holes, all the other sources detected with other messengers had been previously extensively studied through their EM emissions at various wavelengths. Of major recent relevance are the observations in the optical, X-ray and up to 150 keV $\gamma$-rays with the Swift satellite, and between 10 keV X-rays to \mathrel{\mathchoice{\lower 0.5pt\vbox{\halign{\mathnew\displaystyle\hfil#\hfil\cr<\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\textstyle\hfil#\hfil\cr<\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\scriptstyle\hfil#\hfil\cr<\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\scriptscriptstyle\hfil#\hfil\cr<\crcr\sim\crcr}}}}\hbox{~{}TeV} $\gamma$-rays with the Fermi satellite (Fig. 1, upper right) [29], which detected a large number of Gamma-Ray Burst (GRB) sources, active galactic nuclei (AGNs) including blazars, supernovae, etc., as well as a diffuse cosmic $\gamma$-background. Of increasing importance for such sources are the ground-based air Cherenkov imaging telescopes, e.g. MAGIC (Fig. 1, upper middle), HESS, VERITAS [30] and the High Altitude Water Cherenkov observatory HAWC [31, 32], which measure gamma-rays in the 100 GeV to multi-TeV range. These have been amply supported by ground and space observations with multiple radio, infra-red, optical and UV telescopes.

2.2 Developments in Joint Multi-Messenger Astrophysics

$\bullet$  *Solar and Supernova Neutrinos and Photons.- * The two earliest multi-messenger detections involved neutrinos in the MeV range. Davis and collaborators, starting in the 1960’s, detected the electron neutrinos produced by the nuclear reactions that are the energy source for the Sun’s light, using a 600 ton perchlorethylene (cleaning fluid) tank located deep underground in the Homestake mine in South Dakota, US. This neutrino flux was confirmed by various other experiments including the one in the Kamioka mine in Japan, by Koshiba and collaborators. The other early multi-messenger detection was that of neutrinos from a core-collapse supernova, SN 1987a, resulting from inverse beta decay as protons are converted into neutrons. This was detected by three different underground detectors, Kamiokande in Japan, Baksan in the Soviet Union, and Irvine-Michigan-Brookhaven in the US [33, 34, 35]. The neutrino detection preceded significantly the spectacular optical brightening characterizing supernovae. These discoveries earned Davis and Koshiba the Physics Nobel Prize in 2002 [36, 37].

$\bullet$  *Cosmic ray, Gamma-ray and Neutrino Background Interdependences.- * The measurements of the diffuse UHECR energy spectrum by the Pierre Auger Observatory starting in 2008 put on a firm ground the detection of a spectral cutoff above $10^{19.5}$ eV, compatible with the GZK energy losses due to the cosmic microwave background photons [2], after earlier work by HiRes. Then, starting in 2008, the Fermi satellite (following on previous work by COS-B and other missions) measured a diffuse gamma-ray background extending into the sub-TeV range [39]. And starting in 2012-2013 IceCube discovered, with increasing detail, a diffuse high energy neutrino (HEN) background of astrophysical origin at multi-TeV to PeV energies [14, 15]. There is so far no firm identification of the sources of either the UHECR, HEN or gamma-ray diffuse backgrounds, although the extragalactic gamma-ray background is known to be dominated by blazars [40]. However, theoretical relationships and mutual constraints are expected from the basic physics of these three radiations, e.g. Fig. 2 (b). The HENs are produced when UHECRs collide with low energy target photons and nuclei resulting in charged and neutral pions, which decay in a predictable fraction of high energy neutrinos and gamma rays. The resulting energy spectra of neutrinos and photons imply corresponding diffuse backgrounds which must fit the observed results, including also the constraint provided by the observed UHECR background. The fact that the energetics of these three messengers is comparable has led to the idea of a unification of high-energy cosmic particles, e.g., [41, 38]. On the other hand, significant constraints are also placed on generic $pp$ hadronuclear production models of HENs and gamma-rays when when one compares them to the Fermi diffuse $\gamma$-ray flux, especially accounting for the $\gamma\gamma$ cascades initiated by $\gamma$-rays scattering off cosmic radiation backgrounds [42, 43]. The constraints are more stringent for Galactic sources [44]. HAWC [31] is expected to uniquely contribute to measurements of the $\gamma$-ray background in the 10 to 100 TeV energy range, which could strongly constrain the fraction of IceCube neutrinos from Galactic origin. Among $p\gamma$ photomeson production models of HENs, valuable constraints have been put on the contribution of the simpler classical GRB neutrino emission models [45, 46, 47], while leaving open the possibility of contribution by choked GRBs or supernovae driven by choked jets [48, 49, 47].

$\bullet$  *Gravitational Waves and Photons from Binary Neutron Star (BNS) Mergers.- * As the culmination of a long series of previous BNS GW/multi-messenger searches, e.g. [50, 51, 52, 53], the joint GW/EM detection of the transient GW/GRB 170817 was the first high significance proof of the strength of the joint multi-messenger technique in the GW realm [54], e.g. see Fig. 2 (c). The GWs in GW/GRB 170817 showed that this was a neutron star binary merger, providing a measurement of their masses, the distance [55] and gave constraints on the neutron star equation of state [56], while $\gamma$-ray and X-ray measurements by Fermi and Swift showed it was an off-axis short GRB, e.g. [57]. The near simultaneous observation of EM and GW signals from GW170817 showed that they both travel at the speed of light to better than 1 part in $10^{15}$, thereby ruling out many alternative theories of gravity. Optical observations with various telescopes showed that it also manifested itself as a Kilonova, which is an outflow rich in the so-called r-process high atomic number nuclear elements, providing a rich interlocking picture, e.g. [58, 59, 60, 61]. For their role in the discovery of binary gravitational wave sources Barish, Thorne and Weiss received the 2017 Nobel Prize in Physics [62, 63, 64].

$\bullet$  High Energy Neutrinos and Gamma-rays from Blazars.- The joint neutrino [65] and electromagnetic detection [66, 67, 68] of the flaring blazar TXS 0506+056 was an extremely exciting result, being the first time that a known source was shown to be associated (albeit at the $\sim 3\sigma$ level) with a high significance astrophysical high energy neutrino (Fig. 2(d). Blazars are active galactic nuclei (AGNs), which are galaxies with a massive central black hole powering a relativistic jet outflow pointing close to the observer line of sight; they are classified into BL Lac objects and flat spectrum radio-quasars (FSRQs), TXS 0506+056 appearing to be of the BL Lac type 222Recently, however, arguments have been presented [69] indicating that TXS 0506+056 may be an FSRQ instead of a BL Lac object as thought previously.. Blazars are notorious for exhibiting sporadic and intense gamma-ray flaring episodes, one of which was in progress at the time the track-type neutrino was observed. Further analysis indicated that in previous years other neutrinos may have been associated with this source [IC3+18txs0506a]. This provided valuable constraints on the radiation mechanisms and the sources of the diffuse HEN background. Based on simple one-zone emission models where both HENs and gamma-rays originate from the same region, the neutrino is a low probability event [70, 67, 71] and based on a stacking analysis of HENs and blazars it appears that the blazar population as a whole may account for \mathrel{\mathchoice{\lower 0.5pt\vbox{\halign{\mathnew\displaystyle\hfil#\hfil\cr<\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\textstyle\hfil#\hfil\cr<\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\scriptstyle\hfil#\hfil\cr<\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\scriptscriptstyle\hfil#\hfil\cr<\crcr\sim\crcr}}}}10-30\% of the entire IceCube neutrino background [72], so other sources may in any case need to be appealed to.

3 Emerging Questions and Challenges

$\bullet$  The Lack of EM or HEN counterparts of binary black hole mergers is frustrating, with ten binary black hole mergers detected in GWs so far (as of March 2019). Such emissions are expected to be faint at best in BBHs, e.g. [73, 74], but they would be very useful for a better understanding of the binary origin and environment, as well as to get a far better localization than provided by the GWs, e.g. [75, 76, 77, 78]. A much larger sample of BBHs will be needed extending to both smaller and larger masses to test the hypothesis that BBHs provide a cosmologically important dark matter component, e.g. [79, 80, 81].

$\bullet$  Detection of HEN from GW/EM-detected binary neutron stars would provide an example of a “triple-messenger” source, and would clarify major open questions in our understanding of these objects. Possible signals and observing strategies have been discussed in, e.g., [82, 83, 78]. Expected HEN fluxes are low, especially for off-axis jet viewing [84, 85], but in the best case they may be marginally detectable by IceCube (or more plausibly, by a future IceCube-Gen 2), and would greatly aid in clarifying the physics of the relativistic jet and the larger-angle slower outflows which give rise to the GRB, the afterglow, and the kilonova emission of these events.

$\bullet$  Confirmation or refutation of the occurrence of HEN flares in blazars through additional observations of TXS 0506+056 and other AGNs is urgently needed to address the origin of the IceCube background and illuminate any possible connection between the HEN and UHECR backgrounds. Progress in these studies will also require more targeted calculation of AGN neutrino production models, yielding detailed predictions for X-ray and other EM constraints. The stacking analyses of blazar EM flares against observed HENs [72, 86] as well as theoretical arguments [87] indicate that sources other than blazars must provide the dominant contribution to the HEN background, and observational correlation studies involving alternative source candidates may need to be undertaken, e.g. [41].

$\bullet$  The masses and spins of the GW-detected compact mergers offers new puzzles. One is the origin of the ”heavy” binary stellar black holes (\mathrel{\mathchoice{\lower 0.5pt\vbox{\halign{\mathnew\displaystyle\hfil#\hfil\cr>\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\textstyle\hfil#\hfil\cr>\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\scriptstyle\hfil#\hfil\cr>\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\scriptscriptstyle\hfil#\hfil\cr>\crcr\sim\crcr}}}} 30 solar masses), it is not clear how they form and evolve. Another question is why do the LIGO BHs have very low spins or spins mis-aligned with the orbital angular momentum. This is in contrast to X-ray BH candidates, some of which have very large spins. Also the fate of the remnant in the BNS merger GW 170817 is unknown, e.g. how long did the remnant last before turning into a black hole, if it finally did. These and related questions are discussed in, e.g., [27, 88]. Future GW observations could resolve this issue.

$\bullet$  UHECR arrival direction uncertainties are large, and UHECR arrival times are delayed by $\sim$104 to 105 years relative to any simultaneous EM or neutrino emission, so direct correlation attempts have been made only against quasi-steady, non-bursting sources, so far unsuccessfully [2, 89, 90]. At the highest energies, UHECR positional correlations with muon neutrino tracks, UHE neutrinos, and/or gamma-rays could lead to a better pinpointing of the sources. This will require much better instruments and more sensitive neutrino/EM correlation analyses as well as much more detailed production models for likely source candidates.

$\bullet$  Statistically significant measures of UHECR/EM/HEN/GW correlations (or lack thereof) are urgently needed, and it is also necessary to explain the UHECR spectrum and chemical composition with an appropriate distribution of specific sources, e.g. [91, 92]. Observations must be fitted in statistical detail to model predictions of possible candidates, such as AGNs, GRBs, tidal disruption events, clusters of galaxies, etc., and more sophisticated models must be calculated, and tested against the observed diffuse neutrino and gamma-ray backgrounds.

$\bullet$  Theory and simulations are still in their infancy, as far as UHECR, HEN and GW sources. While the HEN/EM inter-relation is in principle straightforward, aside from non-linearities introduced by EM and hadronic cascades, an understanding of the HEN/EM inter-relations with UHECRs is significantly complicated by the fact that the latter are charged, and hence travel in complicated paths, dependent on the magnetic fields, e.g. [93]; in addition, at energies above \mathrel{\mathchoice{\lower 0.5pt\vbox{\halign{\mathnew\displaystyle\hfil#\hfil\cr>\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\textstyle\hfil#\hfil\cr>\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\scriptstyle\hfil#\hfil\cr>\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\scriptscriptstyle\hfil#\hfil\cr>\crcr\sim\crcr}}}}10^{17}-10^{18}\hbox{~{}eV}, cross-section uncertainties start to set in. For BBH and BNS mergers, a lot of progress is urgently required to understand post-merger dynamics, the final state of the remnant, the physics of the ejecta and how BH-NS mergers differ from BNS mergers [94]. Supernova simulations have also been a challenge [95, 96]. Lack of reliable GW waveforms means that we have to rely on sub-optimal techniques for their detection, and it also makes it far more difficult to distinguish between different collapse models/scenarios.

$\bullet$  The ultra-high energy neutrino range $10^{17}-10^{20}\hbox{~{}eV}$ explorations by ANITA [22] and other future experiments need to achieve at least an order of magnitude greater sensitivity to probe the cosmogenic neutrino background. This is due to UHECRs interacting with the cosmic photon backgrounds, and degeneracies are induced by the effect of the UHECR source luminosity function and redshift distribution as well as the CR chemical composition, e.g. [97, 98, 99]. The ANITA anomalous upward-going events [100], if confirmed, are very exciting for what they may tell us about cosmic tau-neutrinos, or possibly about beyond the standard model physics. Concordance studies between ANITA, IceCube and Auger will need to be carried out, together with significantly more detailed theoretical investigations, e.g. [101, 102, 103, 104].

4 Looking Ahead: New Instruments & Results Expected

The spectacular results achieved, mainly in the last decade, by multi-messenger facilities such as those in the first row of Table 1, has opened wide new vistas in high energy astrophysics. This has spurred the building and planning of more sophisticated and more powerful experiments and missions, geared towards the elucidation of the key new questions raised. The second row of Table 1 shows some of the new experiments currently under constructions, while the third row shows some of the next generation of experiments planned for the period between approximately 5 to 15 years from now.

$\bullet$  New Electromagnetic Detectors .- Among the major space-based electromagnetic facilities coming online within the next 5 to 10 years are the Chinese-French Space Variable Object Monitor (SVOM) [105] (Fig. 3(a)), designed for detecting gamma-ray, X-ray and optical transients, scheduled for launch in late 2022. Two other Chinese mission in preparation are GECAM [110], with an all-sky coverage of the sky aimed at detecting GW counterparts in the 6 keV to 5 MeV energy range, scheduled for the mid-2020s; and the time-domain Explorer-class Einstein Probe (EP) [111], with a 3600 sq.deg. field of view sensitivity at 0.5-5 KeV, scheduled for end of 2022. An Israeli-US mission called ULTRASAT [112] has been proposed, with an ultraviolet (250–280 nm), fast slewing ($\sim$ minutes) imaging detector of 250 deg2 field of view, which could detect hundreds of supernovae, $\sim$10 BNS counterparts per year, and $\sim$100 tidal disruption events per year. There is a very strong case to be made for a US X-ray-gamma-ray satellite for providing real-time triggers and data, which would be critical for multi-messenger studies. A possible NASA mission that recently completed Phase A study is the ISS-TAO “Transient Astrophysics Observer” [113], on the International Space Station, with a GTM gamma-ray transient monitor and WFI wide field (350 sq.deg.) lobster-type X-ray imager, whose prime target would be EM counterparts of GW sources, and which might fly by 2032. A significant role in detecting or confirming transients of multi-messenger importance will be played by the ZTF (Zwicky Transient Facility) [114]) and the ASAS-SN facility [115]. Also in the 5 to 10 year timeframe, the multi-national Cherenkov Telescope Array (CTA) [116] (Fig. 3(b)) and the Chinese Large High Altitude Air Shower Observatory (LHAASO) [117] ground-based facilities will survey the sky at TeV-PeV gamma-ray energies, e.g. [30]. Both of these use the air Cherenkov technique, but while CTA includes steerable dishes that provide good angular localization, useful for point sources, LHAASO largely observes as the sky goes by, as does HAWC, which works better for extended or diffuse emission. A major optical/IR survey facility instrument is the Spectroscopic Survey Telescope (LSST) [118], while the Square Kilometer Array (SKA) [119] will provide milliarcsecond spatial resolution images at radio frequencies. Also in preparation are the 30-meter class TMT, ELT and GMT optical/IR ground-based telescopes [120, 121, 122].

For the early 2030s, the European Space Agency ESA is preparing a major flagship X-ray mission called ATHENA [123], which will trace the galaxy formation and metallicity evolution of the Universe with its large area detectors, and can study Population III GRBs. Due for final ESA selection in 2022 for a launch in 2032 is the smaller but nimbler, fast-slewing ($\sim$ minute) satellite THESEUS [124], designed to discover long GRBs at redshifts z\mathrel{\mathchoice{\lower 0.5pt\vbox{\halign{\mathnew\displaystyle\hfil#\hfil\cr>\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\textstyle\hfil#\hfil\cr>\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\scriptstyle\hfil#\hfil\cr>\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\scriptscriptstyle\hfil#\hfil\cr>\crcr\sim\crcr}}}}9 and seek BNS counterparts with a soft X-ray imager, an X- and gamma-ray spectroscopic imager and an 0.7-m class infra-red telescope, which will also provide triggers for ATHENA. The NASA AMEGO satellite [125], sensitive to gamma-rays from 0.2 MeV to \mathrel{\mathchoice{\lower 0.5pt\vbox{\halign{\mathnew\displaystyle\hfil#\hfil\cr>\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\textstyle\hfil#\hfil\cr>\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\scriptstyle\hfil#\hfil\cr>\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\scriptscriptstyle\hfil#\hfil\cr>\crcr\sim\crcr}}}}10 GeV and the German eROSITA [126] 0.3–10 keV space detector will play important roles in in X- and gamma-ray astronomy, as well as in the EM detection of hidden neutrino sources. To complement the above large facilities, it is extremely important to have also various fast, large field-of-view robotic ground-based telescope systems which can follow up transient candidates within seconds after an alert is triggered.

$\bullet$  High and Low Energy Neutrino Detector Improvements and Plans.- High energy neutrino detector planned upgrades include the IceCube High Energy Array and the denser PINGU sub-array, as part of an extended (10 Gtons) IceCube Gen2 [107] (Fig. 3(c)). In the Northern hemisphere, the completion in the Mediterranean sea of the KM3NeT [108] 3 to 4 Gton EU detector is expected by $\sim$ 2026 (Fig. 3(d)); the error box improvements for muon tracks are expected to be \mathrel{\mathchoice{\lower 0.5pt\vbox{\halign{\mathnew\displaystyle\hfil#\hfil\cr<\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\textstyle\hfil#\hfil\cr<\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\scriptstyle\hfil#\hfil\cr<\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\scriptscriptstyle\hfil#\hfil\cr<\crcr\sim\crcr}}}}0.3-0.5 deg2. The relative advantages/disadvantages of ice vs. water as a Cherenkov detector medium are that the light absorption length in ice is $\sim 100$ m vs. $\sim 15$ m for clear ocean water; while the scattering length for ice is $\sim 20$ m vs. $\sim 100$ m for water. Also ocean water contains radioactive 40K, affecting the energy resolution signal to noise ratio. Thus one gets relative better/worse energy resolution, and worse/better angular resolution in ice/water, e.g. [107, 108]. Another Gigaton water-based neutrino detector, Baikal-GVD [127], in the lake Baikal, Russia, is expected by $\sim$ 2021-22. Goals include determining large scale anisotropies and individual source identifications by neutrinos alone or in tandem with other multi-messengers. They will facilitate the use of doublets and multiplets for source population studies, and increase the prospects for identifying galactic sources, reliably identify tau neutrinos, and determine the flavor composition of the high energy neutrino background.

The Hyper-Kamiokande (Hyper-K) [128] next generation megaton water Cherenkov detector, operating at MeV to GeV energies, is located in the Kamioka mine (Japan) and is scheduled to begin construction in 2020. It will be an order of magnitude larger than its predecessor instrument Super-K, where the addition of Gadolinium to the water is providing significantly greater sensitivity. It will be able to detect individual supernova explosions out to $\sim$4 Mpc, roughly one every 3 to 4 years, and in 10 to 20 years it could measure the relict supernova diffuse neutrino flux in the 16–30 MeV energy range [129].

At the highest energies, $10^{17}\hbox{~{}eV}$ to $10^{21}\hbox{~{}eV}$, the ANITA balloon experiment [22] will over the next several years undergo further sensitivity improvements. On a longer timescale of 2022-2032, there is ongoing work and plans for much larger ground-based detectors using the Askaryan radio technique for detecting neutrinos and UHECRs in the same $10^{17}-10^{21}\hbox{~{}eV}$ range, such as ARIANNA [130] and ARA [131], or a possible combination of parts of these efforts (RNO/ARA). These are aimed at detecting the cosmogenic neutrino background component produced by GZK UHECRs, as well as for probing more deeply the nature of the decline of the UHECR spectrum beyond $10^{20}\hbox{~{}eV}$. Another large-scale detector proposal aimed at this goal is the Chinese-led GRAND 10,000 km2 array being planned for the 2025 to 2030s [132], as well as the POEMMA [133] and Trinity [134] projects.

$\bullet$  UHECR Detector Improvements and Extensions.- UHECR facilities undergoing major upgrades include the Auger-Prime addition of 1600 surface scintillation detectors on top of the existing water Cherenkov tanks as well as updated electronics [Auger+17-ICRC]; and the Telescope Array upgrade to the TAx4 configuration four-fold surface enlargement [135]. In the next 5-10 years the planned K-EUSO experiment [136] on the International Space Station (ISS) could achieve uniform exposure across the Southern and Northern hemispheres of $4\times 10^{4}\,{\rm km}^{2}\,{\rm sr}\,{\rm yr}$ per year, an order of magnitude larger than Auger or TAx4. Radio observations with LOFAR (Low Frequency Radio Array) [137] may also contribute substantially to an understanding of UHECRs. In the lower energy range of $10^{12}$ to $>$1015 eV the ISS-CREAM [138] on the International Space Station, building on earlier work of the Cosmic Ray Energetics And Mass (CREAM) balloon flights, as well as the future Chinese mission HERD [139], could greatly increase knowledge about the spectrum and compositions of CR nuclei with charges in the $Z=1-26$ range. In the next 10 to 15 years the planned POEMMA [133] is expected to achieve an increase in the exposure by 100$\times$, while the planned FAST ground array [140] would provide 10x the exposure with high quality events. These advances will address the chemical composition and anisotropy issues of UHECR, the interpretation of the surmised three CR components making up the entire spectrum (e.g. see Fig. 4 of [141]) , the maximum energy of galactic CRs, and will probe in much more detail the nature of the spectral cutoff, the transition from galactic to extragalactic components, the strength of the galactic and intergalactic magnetic fields, etc.

$\bullet$  Gravitational Wave Detectors Improvements & Planned New Facilities.- The upgrades of LIGO and VIRGO are continuing [142], and this will reduce the $90\%$ median credibility error box size for source identifications down to $120-180$ deg2 (with 12-21% with $\leq 20$ deg2) by the 2019+ O3 run. The Japanese KAGRA detector being built in the Kamioka mine in Japan is expected to reach a sensitivity comparable to aLIGO/aVIRGO by 2024. A new LIGO observatory is under construction in India to house the third detector (LIGO-India). With these additional facilities, the expected median 90% localization error box sizes will be 9-12 deg2. The number of expected binary black hole and neutron star detections and the limiting distances are shown in Table 2 (Fig. 5).

Among the next generation of ground-based GW detectors planned for the 2020-30s [143], in the US, the “A+” upgrade to the LIGO facilities has been approved, which should provide a further factor of two increase in detection range beyond the detectors’ advanced sensitivity. A further upgrade, termed LIGO Voyager, would have a new detector operating in the existing LIGO, with cryogenic mirrors in the existing LIGO vacuum envelope. This could bring a further factor of 3 increase in the BNS range (to 1100 Mpc), with a low frequency cut-off down to 10 Hz. In the EU, the planned underground Einstein Telescope [109] (Fig. 3(e)), with three arms of 10 km length, will be able to measure the GW polarization of BBHs and BNS sources from distances 3 to 10 times farther than with the current design sensitivity of the LIGO/VIRGO designs. Further in the future, the Cosmic Explorer ground-based observatory [144] would use 40 km arms to achieve a further order of magnitude improvement in sensitivity over the 10 to 104 Hz frequency range and detect compact binary in-spirals throughout our Hubble volume.

In order to study the merger of the much larger ($10^{6}$ to 10^{8}\mbox{M_{\odot}}) “supermassive” black holes located at the center of galaxies, the EU is planning a large space-based GW detector called eLISA [145], consisting of an interferometer using three small satellites in Solar orbit. A different technique for the exploration of supermassive BH mergers is provided by the Pulsar Timing Arrays (PTAs) [146], such as NANOGrav, PPTA, and EPTA, e.g. [147, 148]. This technique relies on measuring the time delays in the EM radio signals from distant pulsars caused by the space-time variations induced by the GW field of merging BH binaries, and is expected to yield the first stochastic (population-integrated) detections in the near future.

$\bullet$  Exploiting the Multi-Messenger Synergy.- The Astrophysical Multi-messenger Observatory Network (AMON) [149] is a multi-institution consortium which has signed MoUs with a number of observatories using different messengers. One major purpose is to combine disparate rare signals appearing in coincidence, so that even sub-threshold detections in one messenger, when combined with other sub-threshold signals in other messengers, can yield a reliable above-threshold detection. The other purpose is to rapidly distribute interesting transient alerts arising from any observatory to all the other observatories and the community, to facilitate rapid follow-up. The architecture of AMON consists of a central hub with radial spokes connecting to individual observatories, from which it receives individual sub-threshold (and also above threshold) alerts, which are then subjected to analysis and/or redistributed to other observatories. This greatly increases the speed and effectiveness of reaction to a trigger, compared to the large number of traditional individual observatory-to-observatory connections. Observatories that have signed up to the AMON network include, so far, ANTARES, Auger, Fermi LAT, Fermi GBM, HAWC, IceCube, Swift-BAT, VERITAS, and others. Live data streams from IceCube, HAWC, Fermi and Swift are being received by AMON, triggering alerts as in the TXS 0506+056 blazar neutrino and gamma-ray flare discovery [67, 150], and other coincident sub-threshold analyses are being carried out, e.g. for LIGO + Swift BAT, HAWC, and others. The unique and most promising aspect of AMON is its massive emphasis on leveraging multiple live sub-threshold alerts. In addition, sub-threshold analyses are also being carried out using archival data from different individual observatories, e.g. [151]. Other groups are also developing algorithms and strategies for multi-messenger searches, e.g. [152].

$\bullet$  Theory and Simulations.- The high-quality data provided by the facilities outlined above will only yield fruit insofar as it is thoroughly analyzed and interpreted through realistic source models satisfying state-of-the-art physics (while keeping in mind the ultimate possibility of beyond the standard model physics). Such models must be considered at three levels. The basic level is based on an overall conceptual picture, using analytical or semi-analytical source models, with their constitution, dynamics and multi-messenger radiation or micro-physical properties, e.g. [42, 44, 87], see [18, 153, 41, 154, 142] for reviews. The second level involves detailed numerical simulations of the dynamics of the formation of the sources and their evolution leading to the state at which the various types of multi-messenger emission occurs, e.g. [155, 156, 157, 158]. The third level involves detailed calculations and simulations of the radiation physics, using large-scale numerical codes to describe the emission of multi-messengers, followed by their possible changes during propagation from the source to the observer, and their detailed effects on particular types of detectors, e.g. [38, 159, 149, 160, 161, 92]. For low source numbers or low signal rates, diffuse backgrounds must be calculated using simulated source signals convolved over cosmological luminosity and redshift distributions, e.g. [92].

All three of these types of calculations will have to be considerably expanded and refined to address and exploit the potential of the much more detailed data expected from the above new facilities. Even for the semi-analytical studies, the increasing sensitivity and range of the detectors will make it necessary to make use of farther and fainter reaching source luminosity functions, more extensive redshift and mass distributions, intervening plasma and radiation field spectral densities, etc. The source formation and dynamics studies, which are increasingly incorporating general relativistic and magnetohydrodynamic (MHD) effects, will need to be extended to the 3-dimensional regime much more commonly than before, and the use of adaptive mesh and shock capture methods will have to be exploited and developed further. There will be an increasing need for a better cross-calibration of the various Monte Carlo methods used in simulating high energy particle interactions and cascades, incorporating the errors due to theoretical uncertainties, especially in the extrapolation to energies beyond those of accelerator data. As another example, the atomic and nuclear physics of very heavy elements is poorly understood and sparsely studied in the lab, yet to reliably elucidate the sources of r-process elements in the Universe (lately ascribed largely to BNS mergers), the error estimates arising from these theoretical and experimental lab uncertainties will need to be quantified and taken into account. We must emphasize that nuclear and atomic physics lab experiments that can shed light on the r-process details are crucial for our understanding of this important phenomenon.

5 Conclusions and Perspectives

Some of the most important questions that will be addressed in the next 5 to 10 years with upgraded GW detectors such as LIGO, VIRGO and KAGRA, as they improve sensitivity at frequencies above 0.1 kHz, are to explore in detail the lower mass range of binary black hole mergers, to test whether the final outcome of neutron star mergers is a massive neutron star or a black hole, to probe the final ringdown of space-time around a newly formed black hole, to determine the maximum mass of neutron stars, to test whether General Relativity remains valid under extreme density and pressure conditions, and to explore the nature of the central engine of GRBs. As they push towards lower frequencies below 10 Hz it will be possible to probe intermediate mass BHs of 100 to 500 solar masses, important for understanding how the most massive BHs at the center of galaxies form. To get more reliable results, larger interferometer arm lengths such as the $\sim$ 10 km of the Einstein or Cosmic Explorer experiments will be needed. On the 10-20 year timescale, even longer arms, such as the 2.5 million km in the space-based eLISA interferometer, will measure frequencies $\sim 0.1$ Hz which can probe the merger of \mathrel{\mathchoice{\lower 0.5pt\vbox{\halign{\mathnew\displaystyle\hfil#\hfil\cr>\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\textstyle\hfil#\hfil\cr>\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\scriptstyle\hfil#\hfil\cr>\crcr\sim\crcr}}}{\lower 0.5pt\vbox{\halign{\mathnew\scriptscriptstyle\hfil#\hfil\cr>\crcr\sim\crcr}}}}10^{6}\mbox{M_{\odot}} black hole mergers, important for understanding the growth of galaxies and clusters of galaxies in the Universe, and the existence of a primordial GW background left over from the inflationary era of the early Bing Bang.

The neutrino detector upgrades such as PINGU in IceCube and ORCA in KM3NeT will get better statistics in the 1-10 GeV energy range to probe fundamental neutrino physics issues such as the hierarchy of the mass ordering, as well as the mixing angles between flavors, constraining the neutrino masses and testing the existence of sterile neutrinos. These issues are likely to be resolved in the next 5-10 years with the help of these and other detectors, while the next generation IceCube Gen-2 is likely to identify the sources of the origin of high energy neutrino, e.g. [41]. In the Northern hemisphere, the 3 to 4 Gton KM3NeT detector will be able to observe HENs from our galactic center, where most of the (so far undetected) galactic neutrino sources reside. The upgraded ANITA balloons and, if funded, the ARIANNA and RNO/ARA experiments in Antarctica will make progress towards detecting the $10^{17}-10^{19}$ eV cosmogenic neutrinos produced by UHECR protons interacting with the cosmic microwave background, or by UHECR nuclei interacting with the diffuse starlight. These experiments will also complement IceCube-Gen2 in the pursuit of the identification of tau-neutrinos. More reliable determinations may need to wait for larger experiments such as GRAND and POEMMA, which will also address the chemical composition of the highest energy UHECRs. The next Galactic supernova should be an ideal and important event for multi-messenger astrophysics, which can be exquisitely studied by the Hyper-K [128] and JUNO [162] experiments to address fundamental issues of neutrino oscillation and supernova physics, e.g. [163, 164].

The next upgrades of the Auger and TA UHECR arrays are expected to answer the chemical composition question independently of the answers obtained through the above neutrino detectors, providing a much needed consistency check. When the TAX4 array is completed its area will be comparable to Auger’s, and being in the Northern hemisphere while Auger is in the Southern, will prove or disprove possible anisotropies of arrival, also testing whether the UHECR production is dominated by a few nearby sources or more numerous distant ones. Together, both of them will probe the details of the properties of hadronic interactions at energies three orders of magnitude higher than what is achievable in laboratory accelerators. These questions can be more thoroughly investigated from space with K-EUSO and POEMMA in the 10-15 years.

Future all-sky monitors such as the Large Spectroscopic Survey Telescope (LSST) in the optical and the Square Kilometer Array (SKA) in the radio, as well as the Fermi, Swift and expected (2022) SVOM satellites at keV to GeV energies will provide rapid EM triggers and accurate sky localization as well as follow-up capability for studying cataclysmic events such as BH or NS mergers, supernovae, gamma-ray bursts, AGN flares, etc., where CRs, neutrinos and GWs are also expected in varying amounts. The strength and mix of these different messengers is model-dependent, and multiple triggers in different messengers, as well as model development and extensive simulations, will be the key for understanding the physics of these energetic sources.

The key to fully exploit the power of these new facilities is the multi-messenger approach, due to the complementary advantages and limitations of the different messenger particles. Cosmic rays provide unique information about particle accelerators well beyond terrestrial laboratory capabilities, and about source magnetic fields and total energetics, but they do not reach us from beyond $\sim$100 Mpc and have at best poor angular resolution. Neutrinos reach us from the most distant reaches of the Universe, and probe the inner, denser regions of the most energetic and cataclysmic events, and ultra-high energy neutrinos, being intimately linked to UHECRs, can provide unique clues as to how the latter reach their enormous energies. Gravitational wave observations probe the most compact regions of high energy sources; the GW wave strain amplitude at earth is directly proportional to the source compactness, measured in terms of $GM/c^{2}R$, and the GW luminosity goes as the fifth power of the compactness. GWs provide excellent information about central object masses, angular momenta, and distances, and they will eventually be detected from the farthest distances and earliest epochs in the Universe, but they have modest angular angular resolution at best, and do not probe the bulk of the stellar or baryonic mass of their sources. Electromagnetic waves provide excellent angular resolution, velocity and redshift determination capabilities, but the opacity of matter prevents them from probing the inner, denser regions of astronomical sources, while at higher gamma-ray frequencies they cannot reach us directly from the much grater distances probed by neutrinos or gravitational waves. By using several messengers in conjunction, astronomers and physicists can hedge the foibles of each against the advantages of the others, forging them into a formidable toolkit for probing the highest energies and densest, most violent corners of the Universe, and in this fashion, putting our physical theories of the universe to their most extreme and exacting possible tests.

6 Explanatory Boxes

**BOX 1: Neutrino Production and Oscillations

**Cosmic ray protons $p_{cr}$ interacting with target protons $p_{t}$ and target photons $\gamma_{t}$ lead to a reduced energy proton $p$ or neutron $n$, and to intermediate charged or neutral short-lived unstable particles, such as pions $\pi^{\pm,0}$, muons $\mu^{\pm}$, neutrons $n$ and (at higher energies) Kaons $K^{\pm}$, whose decay results in neutrinos $\nu_{i}$ of different flavors $i$, $\gamma$-rays and electrons or positrons $e^{\pm}$. CR nucleons, on the other hand, are primarily subject to spallation against target protons and photo-disintegration against target photons, both resulting in smaller nucleons including protons, the latter subsequently undergoing the same above mentioned interactions.

[TABLE]

[TABLE]

[TABLE]

The primary CR proton’s mean relative energy loss per interaction, called the inelasticity, is $\kappa_{pp}\sim 0.5$ for $pp$ (or $pn$) and $\kappa_{p\gamma}\sim 0.2$ for $p\gamma$ interactions. For $p_{cr}$ interactions with target protons (or target neutrons) the mean ratio of charged to neutral pion secondaries is $r_{\pm/0}\sim 2$, and for interactions with target photons it is $r_{\pm/0}\sim 1$. The electrons and positrons $e^{\pm}$ quickly lose their energy via synchrotron or inverse Compton scattering, resulting in further $\gamma$-rays, so the final result of the $p_{cr}+p_{t}/\gamma_{t}$ interactions are high energy neutrinos $\nu_{i}$ and $\gamma$-rays. The mean energy of the resulting neutrinos and $\gamma$-rays is $\sim 0.05$ and $\sim 0.1$ of the initial CR proton’s energy.

Once neutrinos of any flavor are produced, during their travel from the source to the observer the neutrinos of any flavor can change into neutrinos of any of the three flavors, in the so-called neutrino oscillation phenomenon. As a result, independently of the ratio of neutrino flavors initially produced a the source where the CR interactions took place, after traveling over astronomical distances typically all three neutrino flavors are present at the observer. For the neutrino sources considered here, the observable neutrino flavors are expected to be oscillation-averages due to the large source distances and the finite energy resolution of the neutrino observatories.

**BOX 2: Multi-messengers and their inter-relation  **

A multi-messenger source might emit two, three, or even all four different types of messengers.

For instance, panel (G) of Fig.6 shows a binary neutron star merger such as the GW/GRB 170817 event, from which two types of multi-messengers, gravitational waves (GW) and photons ($\gamma$), were observed [54, 57, 59]. Such sources may also emit high energy neutrinos (HENs) and cosmic rays (CRs) e.g. [84, 85, 165], although for this particular source theories predict fluxes too low for current detectors; if so, even closer events or next-generation HEN facilities will be required to observe HEN from these sources. Another panel, (B), shows a tidal disruption event (TDE) of a star by a massive black hole; in this case shocks in the disrupted gas can accelerate particles and lead to CRs and HENs, e.g. [166, 167, 168, 169]. TDEs involving white dwarf stars and $\sim$1000 $M_{\odot}$ black holes lead to strong low-frequency ($\sim$1 mHz) gravitational wave emission that could be observed by the forthcoming eLISA mission.

A solitary supermassive black hole with a jet may emit gamma-rays, HEN, and cosmic rays (panel B), as we suspect occurred during the 2017 flaring episode of the BL Lac-type blazar TXS 0506+056 [66, 67, 68, 65, 70, 71]. Here and in related sources, the coproduction of CRs, HEN, and high-energy gamma-rays is anticipated, as the physics of these three messengers are closely connected – high-energy particle acceleration and shocks lead to the interaction of highly-relativistic protons (or nuclei) with ambient gas or intense radiation fields, resulting in neutrinos, gamma-rays, and $e^{\pm}$. For single objects, even those of extreme mass and undergoing substantial accretion, relatively weak gravitational wave emission is expected as the time-varying quadrupole moment (which requires the breaking of azimuthal symmetry) in these cases are thought to be minimal. The sole exception would be a Galactic engine-driven supernova or a Galactic supernova (panels D and E), which would be sufficiently nearby that detection of coherent or incoherent gravitational waves by current and future ground-based detectors is anticipated. As far as thermal ($\sim 10$ MeV) neutrino detection from such supernovae, IceCube is well matched, having for such emission also a roughly galactic reach. A challenge for theory is to predict the amplitude and spectrum of GW and neutrinos from different types of supernovae.

Strong GW emissions have been observed from the mergers of compact binary systems, either from two merging stellar-mass black holes (panel F) [27], two merging neutron stars (panel G) [54], or (in the future) BH-NS mergers, because the final in-spiral to coalescence yields a strong gravitational wave signal in the “sweet spot” frequency range for ground-based gravitational wave detectors. In the case of 30 $M_{\odot}$ + 30 $M_{\odot}$ black hole binary systems, such coalescence events can already be observed out to $\sim$500 Mpc distances [142]. However, in the case of BH-BH mergers little EM flux is expected, because the ambient matter density (protons, electrons) in the vicinity of the binary, at the time of the merger, is typically very low. A key exception is the accreting supermassive black holes at the centers of massive galaxies, which are expected to merge in the wake of the coalescence of their component galaxies. These SMBH mergers are key targets for the eLISA mission, and may well exhibit accompanying EM, CR, and HEN emission [170].

Acknowledgments: We are grateful to Stephane Coutu, Douglas Cowen, Miguel Mostafá and Bangalore Sathyaprakash and the referees for useful discussions and comments.