AGM2015: Antineutrino Global Map 2015

Every second greater than 1025 antineutrinos radiate to space from Earth, shining like a faint antineutrino star. Underground antineutrino detectors have revealed the rapidly decaying fission products inside nuclear reactors, verified the long-lived radioactivity inside our planet, and informed sensitive experiments for probing fundamental physics. Mapping the anisotropic antineutrino flux and energy spectrum advance geoscience by defining the amount and distribution of radioactive power within Earth while critically evaluating competing compositional models of the planet. We present the Antineutrino Global Map 2015 (AGM2015), an experimentally informed model of Earth’s surface antineutrino flux over the 0 to 11 MeV energy spectrum, along with an assessment of systematic errors. The open source AGM2015 provides fundamental predictions for experiments, assists in strategic detector placement to determine neutrino mass hierarchy, and aids in identifying undeclared nuclear reactors. We use cosmochemically and seismologically informed models of the radiogenic lithosphere/mantle combined with the estimated antineutrino flux, as measured by KamLAND and Borexino, to determine the Earth’s total antineutrino luminosity at . We find a dominant flux of geo-neutrinos, predict sub-equal crust and mantle contributions, with ~1% of the total flux from man-made nuclear reactors.


INTRODUCTION
The neutrino was proposed by Wolfgang Pauli in 1930 to explain the continuous energy spectrum of nuclear beta rays.By Pauli's hypothesis the missing energy was carried off by a lamentably "undetectable" particle.Enrico Fermi succeeded in formulating a theory for calculating neutrino emission in tandem with a beta ray 1 .Detecting Pauli's particle required exposing many targets to an intense neutrino source.While working on the Manhattan Project in the early 1940s Fermi succeeded in producing a self-sustaining nuclear chain reaction, which by his theory was recognized to copiously produce antineutrinos.Antineutrino detection projects were staged near nuclear reactors the following decade.In 1955, Raymond Davis, Jr. found that reactor antineutrinos did not transmute chlorine to argon by the reaction: 37 Cl ( νe , e − ) 37 Ar 2 .This result permitted the existence of Pauli's particle only if neutrinos are distinct from antineutrinos.Davis later used the chlorine reaction to detect solar neutrinos using 100,000 gallons of dry-cleaning fluid deep in the Homestake Gold Mine.Reactor antineutrinos were ultimately detected in 1956 by Clyde Cowan and Fred Reines by recording the transmutation of a free proton by the reaction 1 H ( νe , e + ) 1 n 3,4 .This detection confirmed the existence of the neutrino and marked the advent of experimental neutrino physics.
Almost 60 years later neutrino research remains an active and fruitful pursuit in the fields of particle physics, astrophysics, and cosmology.In addition to nuclear reactors and the Sun, detected neutrino sources include particle accelerators 5 , the atmosphere 6,7 , core-collapse supernovae [8][9][10][11] , the Earth 12,13 , and most recently the cosmos 14 .We now know that neutrinos and antineutrinos have "flavor" associations with each of the charged leptons (e, µ, τ ) and these associations govern their interactions.Neutrino flavors are linear combinations of neutrino mass eigenstates (ν 1 , ν 2 , and ν 3 ).This quantum mechanical phenomenon, known as neutrino oscillation, changes the probability of detecting a neutrino in a given flavor state as a function of energy and distance.Neutrino flavor oscillations along with their low cross section provide a glimpse into some of the most obscured astrophysical phenomena in the universe and most recently the otherwise inaccessible interior of our planet.Antineutrinos emanating from the interior of our planet constrain geochemical models of Earth's current radiogenic interior.Antineutrino observations of the modern Earth's interior coupled with cosmochemical analysis of chronditic meteorites from the early solar system allow scientists to model the geochemical evolution of the Earth across geologic time.timing (∼50ps) multichannel plates 17 coupled with Gd/Li doped scintillator will allow for the imaging of antineutrino interactions.The imaging and subsequent reconstruction of antineutrino interactions produce directionality metrics.Directionality information can be leveraged for novel geological investigations such as the geo-neutrinographic imaging of felsic magma chambers beneath volcanos 18 .These exciting geophysical capabilities have significant overlap with the non-proliferation community where remote monitoring of antineutrinos emanating from nuclear reactors is being seriously considered 19 .
Antineutrino Global Map 2015 (AGM2015) shown in Figure 1 merges geophysical models of the Earth into a unified energy dependent map of νe flux, both natural and manmade, at any point on the Earth's surface.We provide the resultant flux maps freely to the general public in a variety of formats at www.ultralytics.com/agm2015.AGM2015 aims to provide an opensource infrastructure to easily incorporate future neutrino observations that enhance our understanding of Earth's antineutrino flux and its impact on the geosciences.In this study we first describe the particle physics parameters used in propagating antineutrino oscillations across the planet's surface as shown in Figure 2. A detailed description of the incorporation of anthropogenic and geophysical neutrino energy spectrum from 0-11 MeV is presented which allows for the four-dimensional generation (latitude, longitude, flux, and energy) of the antineutrino map as shown in separate energy bins in Figure 3.A vertically stratified model of the Earth's density, shown in Figure 4, based on seismological derived density models are combined with a cosmochemical elemental abundances to determine the geological signal of antineutrinos.This signal is then constrained by geo-neutrino measurements from KamLAND and Borexino and first order uncertainties associated with AGM map are then presented.

NEUTRINO OSCILLATIONS
AGM2015 incorporates the known 3-flavor oscillation behavior of antineutrinos.This starts with the standard 3-flavor Pontecorvo Maki Nakagawa Sakata (PMNS) matrix U : where c ij = cos (θ ij ) and s ij = sin (θ ij ), and θ ij denotes the neutrino oscillation angle from flavor i to flavor j in radians.In this paper we assume θ 12 = 0.587 +0.019 −0.017 per a global fit by Fogli et al. 20 in the case of θ 12 , and by measurements at the Day Bay experiment 21 in the case of θ 13 .Phase factors α 1 and α 2 are nonzero only if neutrinos are Majorana particles (i.e. if neutrinos and antineutrinos are their own antiparticles), and have no influence on the oscillation survival probabilities, only on the rate of possible neutrino-less double beta decay.We assume α 1 = α 2 = 0 in this work.We likewise assume phase factor δ = 0, though this assumption may change in the future if evidence is found to support neutrino oscillations violating charge parity (CP) symmetry.
The probability of a neutrino originally of flavor α later being observed as flavor β is: where E is the neutrino energy in GeV, L is the distance from its source the neutrino has traveled in km, and the delta-mass term ∆m The last approximation assumes no charge parity (CP) violation (δ = 0), causing the imaginary terms to fall out.The * symbol denotes a complex conjugate, and U ij denotes the element of the PMNS matrix U occupying the i th row and j th column.Equation 3 can be employed to determine the 'survival probability' of a νe of energy E GeV later being observed as the same flavor a distance L km from its source.In particular, the P e→e survival probability of most interest to this paper can be expressed as Equation 4: For simplicity we ignore the Mikheyev-Smirnov-Wolfenstein (MSW) effect 22 on neutrinos as they travel through the Earth.We use the neutrino mixing angles and mass constants from Fogli et al. 2012 20 and Daya Bay 2014 21 to evaluate Equation 4 for all source-observer ranges and energies used in AGM, giving us the survival probability of seeing each source from each point in the map at each energy level.This is not a trivial task, requiring > 1 × 10 15 evaluations of Equation 4 for a full AGM2015 rendering.This is broken down into ∼ 1 × 10 6 point sources, ∼ 1 × 10 6 locations on the map at which the flux is evaluated, and ∼ 1 × 10 3 energy bins spanning the 0 < E νe < 11MeV energy range as shown in Figure 3. Equation 4 can best be visualized in Figure 2, which shows the θ 12 ripples in the 3MeV worldwide reactor-ν e flux.

REACTOR ANTINEUTRINOS
Reactor νe experiments have proven the viability of unobtrusive reactor monitoring and continue to contribute important information on neutrino properties including the possibility of additional light "sterile" neutrino flavors 23 .We use the International Atomic Energy Agency's (IAEA) Power Reactor Information System (PRIS) 24 to identify and locate 435 known man-made reactor cores in operation at the time of this writing.PRIS categorizes reactors into four states: AGM2015 includes all "Operational" or "Temporary Shutdown" reactors, including many reactors in Japan affected by the Fukushima-Daiichi disaster, which are classified as "Temporary Shutdown" rather than "Permanent Shutdown."PRIS shows 435 Operational and Temporary Shutdown reactor cores distributed among 193 sites, with 870 GW th total output after load factor considerations, and 72 reactor cores among 42 sites (total 156 GW th at 100% load factor) currently Under Construction.The PRIS database reports thermal capacity directly, which is typically about three times electrical capacity (most reactors are about 30% efficient in converting heat into electricity).Historical "load factors" of each core are used to convert the total thermal capacity to projected current and future thermal power output.Load factors account for the down-time related to maintenance and other outages, allowing for AGM2015 to be a reliable estimate of worldwide antineutrino flux in its year of release.Typical PRIS load factors range from 70% to 90%.The 3MeV AGM2015 reactor-ν e flux map due to these 435 Operational and Temporary Shutdown reactor cores is shown in Figure 2.
The reactor-ν e energy spectrum assumes the shape of an exponential falloff in energy, with most reactor-ν e released at the lowest energies.To obtain an νe /s/GW th reactor scaling we assume: • The mean energy released per fission E f is around 205 MeV 25 .
• The reactor thermal power P th is related to the number of fissions per second N f = 6.24×10 21 • 6 νe created per fission 19 • 2 νe created per fission on average above the inverse beta decay (IBD) energy threshold of E ν ≥ 1.8 MeV 19 .
These assumptions yield 1.83×10 20 νe /s/GW th (of all energies) emitted from a reactor, with 0.61×10 20 νe /s/GW th emitted above the 1.8 MeV IBD detection threshold.This is a mean value representative of a typical pressurized water reactor at the beginning of its fuel cycle 19 .We find the summed worldwide reactor power output to be 870GW th , and the resultant νe luminosity to be 1.6 × 10 23 νe /s and 0.04TW.Jocher et al. 26 is recommended for a deeper discussion of reactor-ν e detection via IBD detectors.

GEONEUTRINOS
Observations from geology, geophysics, geochemistry, and meteoritics allow for a range of non-unique solutions for the composition of the Earth.The relative proportion of Fe, O, Mg, and Si in chondritic meteorites individually varies by ∼ 15% each and reflects spatial and temporal differences in where these rocks formed in the early solar nebula.Likewise, refractory elements have 25% variation in their relative abundance, which translates into a factor of two in absolute concentration difference of these elements.Even greater enrichment factors of these elements occur when the volatile inventory (e.g., H 2 O, CO 2 , N 2 ) is mostly lost, as during terrestrial planet assembly.Finally, because the Earth's core is taken to have negligible amounts of Th, U and K [27][28][29] , due to their limited solubility in core-forming metallic liquids, this becomes another 50% enrichment factor in the radiogenic elements in the silicate Earth.Consequently, compositional models predict between 10 and 30 ng/g U (and Th/U = 3.9, the chondritic ratio) for the silicate Earth.Given the planetary ratio of Th/U and K/U (1.4 × 10 4 ) 30 , and the absolute U content of the silicate Earth, its heat production for a 10 ng/g U model roughly corresponds to a surface heat flow of 10 TW and likewise 30 ng/g U to ∼ 30 TW. Estimates of the Earth's radiogenic heat production thus vary from low power models (10-15 TW of power from K, Th, and U), through medium power models (17-22 TW), and to high power models (> 25 TW) 31 .Accordingly, detecting the Earth's flux of geoneutrinos can provide crucial data to test competing theories of the bulk Earth.
Two observatories, one in Japan (KamLAND) and one in Italy (Borexino), are making ongoing measurements of the surface flux of geoneutrinos at energies above the IBD threshold energy E ν ≥ 1.8 MeV.At Japan the flux measurement is (3.4 ± 0.8) × 10 6 cm −2 s −112 , while at Italy the flux measurement is (4.that it is sometimes convenient to express geoneutrino flux as a rate of recorded interactions in a perfect detector with a given exposure using the Terrestrial Neutrino Unit (TNU) 32 , however in this work we focus on simple νe flux (ν e /cm 2 /s) and luminosity (ν e /s).AGM models the Earth as a 3D point cloud consisting of roughly 1 million points.National Oceanic and Atmospheric Administration (NOAA) Earth TOPOgraphical 1 (ETOPO1) "ice" data 33 is used to provide worldwide elevations with respect to the World Geodetic System 84 (WGS84) ellipsoid.Zero-tide ocean surface corrections to the WGS84 ellipsoid were obtained from the National Geospatial-Intelligence Agency (NGA) Earth Gravitational Model 2008 34 (EGM2008) for modeling the ocean surface elevations around the world.Underneath these surface elevations we model 8 separate crust layers using CRUST 1.0 35 , shown in Figure 4, as well as a 9th adjoining layer per Huang et al. 36 which reaches down to the spherical mantle, creating a seamless earth model.Certain crust tiles which are too large (about 200 km across at the equator) to be adequately modeled as point sources are instead modeled as collections of smaller tiles using numerical integration, which recursively subdivides large tiles into progressively smaller sub-tiles until the contribution of each is less than 0.001 TNU.
Geoneutrino flux is produced from the decay of naturally occurring radioisotopes in the mantle and crust: 238 U, 232 Th, 235 U, 40 K, 87 Rb, 113 Cd, 115 In, 138 La, 176 Lu, and 187 Re 37 .However, we only consider 238 U and 232 Th in our flux maps as all other elements' energy spectrum is considerably below the IBD energy threshold of E ν ≥ 1.8 MeV.All abundances for the crust and mantle can be seen in Table II.As shown in Table III, K is the largest contributor to νe luminosity but its energy is below the IBD threshold.All elements other than U, Th, and K have a negligible contribution to the Earth's νe luminosity.
Successful detection of νe below 1.8 MeV remains elusive; if successful the incorporation of the remaining radioiso-    30 , Gando et al. 12 , and Bellini et al. 13 .
topes would be beneficial to future versions of AGM.The Earth's core was assumed to have no significant contribution to the νe flux due to limiting evidence for a georeactor 38 and no appreciable amount of 238 U, 232 Th, or 40 K isotopes 27 .While certain core models support upper limits of K content at the ∼100 ppm level 28 , which would be sufficient for up to ∼1-2TW of radiogenic heating in the present day, "constraints on K content are very weak" 29 , and in the absence of stronger evidence we've chosen to assume a K-free core.
Mantle abundances were derived from empirical geo-neutrino measurements at KamLAND 12 and Borexino 13 .We deconstructed the reported geo-neutrino flux from each observation into separate contributions from U and Th according to a Th/U ratio of 3.9.From each of these, we subtracted the predicted crust flux contributions 36 at each observatory, averaging the asymmetric non-gaussian errors, to arrive at estimates of the mantle contributions.We then combined the estimates of the mantle U flux and the mantle Th flux contributions from each observation in a weighted average.The resulting best estimates for the mantle U and Th flux contributions were finally converted to homogeneously distributed mantle abundances using the spherically symmetric density profile of the Preliminary Reference Earth Model (PREM) 39 along with a corresponding correction to account for neutrino oscillations.Corresponding values for K were found by applying a K/U ratio of 13, 800 ± 1300 30 .The resulting AGM2015 U, Th and K mantle abundances are presented in Table II.The main sources of uncertainty in these estimates are the observational errors in the flux measurements and limited knowledge of the subtracted crust fluxes.A detailed description of the methods and relevant conversion factors used here are presented in Dye 31 .AGM2015 neutrino luminosities are for total numbers of neutrinos.Although almost all are originally emitted as electron antineutrinos, on average only ∼0.55 of the total remain so due to neutrino oscillations.We calculate the total Earth νe luminosity to be 3.4 +2.3 −2.2 × 10 25 νe s −1 .A detailed breakdown of 238 U, 232 Th, and 40 K geoneutrino luminosity from the lithosphere and mantle can be seen in Table III (for all energies), as well as in Table IV for E ν ≥ 1.8 MeV. Figure 5 shows the combined AGM2015 crust+mantle νe flux.
A comparison of panels in Figure 4 reveals a more luminous upper crustal signature for the Tibetan plateau, Himalayan and Zagros Mountains, and the North American Cordillera.Whereas the lower crust of the central Andes Mountains is more luminous compared to the aforementioned bright regions of the upper crust which on average are less luminous.One potential explanation for contrasting observations might be repeated upper crustal thrusting in the Himalayas and like regions, while magmatic additions to the lower crust of the Andean arc could account for their relative brightness.These findings deserve further investigation to understand better their origin and distribution, particularly from seismic and geological perspectives.

UNCERTAINTY
The underlying interior structure and composition of the Earth is, in some regards, still poorly understood.The concentration and distribution of radioisotopes, whose decay chains produce geoneutrino flux, dominate the uncertainties.Therefore modeling of the distribution, energy spectra, and total flux of geoneutrinos remains a challenging task on its own.A full description of the uncertainty in each element of the AGM flux maps is not available at this time, however, we have defined the uncertainties in the specific building blocks of the AGM in Tables I and II as well as the systematic uncertainties present in the various geoneutrino luminosity categories in Tables III and IV.The uncertainties in Tables I and II in particular can be used to create Monte Carlo instances of the AGM flux maps, which could be used to evaluate the variances in each element of the AGM flux map, as well as co-variances between map elements.Such a full-scale Monte Carlo covariance matrix is impractical, however, due to the large number of map elements, ∼300 Million, which would end up producing a 300 Million × 300 Million full-size covariance matrix.
In most parts of the AGM2015 map uncertainties are strongly correlated over space and energy.This is due to the fact that the greatest uncertainty lies in ingredients that affect all map elements nearly equally, such as elemental abundances in the Earth's crust and mantle.Other ingredients which might introduce more independent uncertainty, such as the volume or density of specific crust tiles, are much more likely to introduce minimal correlations, or only slight regional correlations.Since reactor-ν e flux is generally better predicted than geo-ν e flux, and since near a core the reactor-ν e flux will dominate the overall νe flux, we would expect smaller fractional uncertainties near reactors than in other regions of the world.
We attempt to apply systematic uncertainties to AGM2015 where appropriate, and to derive these uncertainties from previous work in the field where possible rather than reinvent the wheel.For our uncertainty models we turn to Huang et al. 36 and Dye 31 .Doing so allows us to apply systematic uncertainties to the various νe source categories in Tables III and IV while avoiding the significant computational burden of a full Monte Carlo analysis, as well as the questions that would arise afterward of how to describe the various levels of regional correlation in flux uncertainty across space and energy.

CONCLUSION
Electron antineutrino measurements have allowed for the direct assessment of 7-29TW 31 power from U and Th along with constraining a geo-reactor < 3.7 TW at the 95% confidence level 12 .Such measurements promise the fine-tuning of BSE abundances and the distribution of heat-producing elements within the crust and mantle.Several such models of the Earth's antineutrino flux 32,40,41 existed before the observation of geoneutrinos; with several recent models being presented with the inclusion of geoneutrinos 36,42,43 .All of the aforementioned models incorporate several geophysical models based on the crust and mantle from traditional geophysical measurements (seismology, chondritic meteorites, etc.)This effort, AGM2015, aims to consolidate all these models into a user-friendly interactive map, freely available to the general public and easily accessible to anyone with a simple web browser at www.ultralytics.com/agm2015.
Future work includes completing a more detailed uncertainty study using Monte Carlo methods.Such a study requires an accurate understanding of the uncertainty in each of the AGM elements listed in Tables I and II, however, as well as the correlations that govern their interactions.Future geo-ν e measurements, updated flavor oscillation parameters, advances in crust/mantle models, and the ongoing construction and decommissioning of nuclear reactors around the world necessitates a dynamic AGM map capable of changing with the times.For this reason we envision the release of periodic updates to the original AGM2015, which will be labeled accordingly by the year of their release (i.e."AGM2020").METHODS AGM2015 uses CRUST1.0 35 to model the Earth's crustal density and volume profile via eight stratified layers.Elemental abundances for U, Th and K, and isotopic abundances for 238 U, 232 Th and 40 K for each layer were defined by Huang et al. 36 .These values were coupled to well known isotope half-lives and multiplicities to create νe luminosities emanating from each crust tile.A similar approach was taken with the Earth's mantle, with elemental abundances derived via estimates of geo-ν e flux at KamLAND and Borexino 44 and density profiles supplied via PREM 39 .
Man-made reactors were modeled via the IAEA PRIS 24 database, with reactor-ν e luminosities found to scale as 1.83 × 10 20 νe /s/GW th .Reactor-ν e spectra were modeled as exponential falloffs from empirical data 19 , while 238 U, 232 Th and 40 K spectra were modeled based on the work of Sanshiro Enomoto 37 .
Luminosities from each point-source were converted to fluxes at each map location via the P e→e survival probability shown in Equation 4, and a full understanding of the source spectra of each point-source enabled a complete reconstruction of the observed energy spectra at each map location."Smart" integration was applied where necessary to more accurately portray crust and mantle tiles as volume-sources rather than point-sources.All modeling and visualization was done with MATLAB 45 .Google Maps and Google Earth multi-resolution raster pyramids created with MapTiler 46 .All online content available at www.ultralytics.com/agm2015.

FIG. 5 .
FIG. 5. AGM2015 geoneutrino flux due to 238 U and 232 Th decay in the Earth's crust and mantle.Flux units are νe/cm 2 /s at the Earth's surface.Map includes νe of all energies.Figure created with MATLAB 45 .
TABLE II.AGM2015 distribution and properties of U, Th, and K, which are the main emitters of electron antineutrinos.Abundances in the various stratified crustal layers are shown at the top of the table, including Oceanic Crust (OC), Continental Crust (CC), and Lithospheric Mantle (LM).Relevant isotopic properties are presented at the bottom of the table.Note the abundances are unit-less fractions.Uncertainties shown in this table derived from Huang et al. 36 , Arevalo et al. 30 , Gando et al. 12 , and Bellini et al. 13 .

TABLE IV .
Contribution of geoneutrino luminosities L in AGM2015 above the IBD threshold Eν ≥ 1.8 MeV for 238 U, 232 Th, and 40 K νe emitted by the Earth.The reactor-νe luminosity ≥ 1.8 MeV is 0.6 ± 0.1 × 10 23 νe/s and 26 GW for the world's 435 reactor cores, which together output 870 GW th 24 .Uncertainties shown in this table derived from Huang et al. 36 , Arevalo et al.