ANTARES and IceCube combined search for neutrino point-like and extended sources in the Southern Sky

Giulia Illuminati (for the ANTARES; IceCube Collaborations)

REVIEW 2 major objections 5 minor 12 references

ANTARES and IceCube combined search for neutrino point-like and extended sources in the Southern Sky

T0 review · 2 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Combining ANTARES and IceCube data improves Southern-Sky neutrino-source sensitivity by a factor of about two, yet the analysis finds no significant source.

desk verdict First combined ANTARES+IceCube Southern-sky search: a clean null result with a plausible factor-two sensitivity gain, delivered as an honest but notationally sloppy proceedings. read the letter →

arxiv 1908.07439 v1 pith:FCNAADVH submitted 2019-08-20 astro-ph.HE

Giulia Illuminati (for the ANTARES , IceCube Collaborations) This is my paper · ORCID

classification astro-ph.HE

keywords neutrinoastronomypoint-likesourcesextendedSouthernSkymaximumlikelihoodANTARESIceCubefluxupperlimits

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper asks whether combining the data of two very different neutrino telescopes—IceCube at the South Pole and ANTARES in the Mediterranean Sea—can improve the search for cosmic neutrino sources in the Southern Sky. It reports a sensitivity gain of roughly a factor of two over either experiment alone, but no discovery: the joint maximum-likelihood scan finds no cluster of events that stands out from the atmospheric-neutrino background. The most significant Southern-Sky point-like excess has a post-trial probability of 0.18, and the most significant result, a 2-degree extended excess near the Galactic Centre, has a post-trial probability of 0.03 (about 1.9 sigma in the one-sided convention). The paper therefore establishes that, in these data, the Southern-Sky neutrino sky is consistent with background, and it reports 90% confidence-level upper limits on several classes of hypothetical sources.

What carries the argument

The engine of the analysis is a joint unbinned maximum-likelihood fit over seven event samples: ANTARES tracks and showers plus the IceCube runs IC40, IC59, IC79, and IC86. The likelihood multiplies, for each sample and event, a weighted mixture of signal and background probability density functions, each formed as a product of a spatial and an energy term. For IceCube the spatial signal PDF is a two-dimensional Gaussian with effective width $\sigma_{\mathrm{eff}} = \sqrt{\sigma^2 + \sigma_s^2}$; for ANTARES it is a Monte-Carlo-derived point-spread function, with $\sigma_s$ encoding a Gaussian source extension. The test statistic $Q = \log L(\hat{n}_s, \hat{\gamma}) - \log L(n_s = 0)$ is scanned over the sky in $1^\circ \times 1^\circ$ bins. Trial-corrected p-values come from comparing the observed $Q$ with the distribution obtained from background-only pseudo-experiments in which event times are randomized to destroy clustering.

What would settle it

Re-run the identical pipeline on pseudo-experiments that randomize event directions rather than event times; if the tail of the resulting test-statistic distribution is heavier than the time-randomized one, the reported post-trial p-values (0.18 full-sky, 0.03 Galactic Centre) would overstate the significance of the observed excesses. A complementary check is to inject a simulated $E^{-2}$ point-source flux at the level of a reported 90% upper limit at a random Southern-Sky position and verify that the analysis recovers it at the claimed sensitivity.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central result is a null result obtained with an improved tool: none of the five searches—the full Southern-Sky scan, a list of 57 candidate sources, the Galactic Centre region, Sagittarius A*, and the supernova remnant RXJ 1713.7-3946—produces evidence of a cosmic neutrino source at the 90% confidence level. The best full-Sky point-like cluster yields 5.7 fitted signal events with a spectral index of 2.5 and a post-trial p-value of 0.18. The Galactic Centre extended-source search with $\sigma_s = 2.0^\circ$ gives 20.3 fitted events, a spectral index of 3.0, and the smallest post-trial p-value in the paper, 0.03. The constructive claim is that the combined analysis, exploiting ANTARES' better angular resolution at low energies in the Southern Sky alongside IceCube's larger effective volume, reaches roughly twice the sensitivity of either standalone search, as quantified by the 90% C.L. median sensitivity curves.

Load-bearing premise

The quoted significances and sensitivity gain assume that the time-randomized background-only pseudo-experiments reproduce the true joint atmospheric background of the two detectors, including its energy and directional distributions.

Editorial extensions

If this is right

A Southern-Sky neutrino source whose $E^{-2}$ flux lies at or above the combined 90% C.L. upper limits would be excluded by this analysis, so the limits set the benchmark for future hadronic-emission models of these objects.
The smallest post-trial p-value in the paper occurs in the 2-degree extended Galactic Centre search (0.03), making that region and source-size hypothesis the one most likely to yield a real signal with more data.
Because the two data sets enter a single likelihood, the same procedure can be applied without structural changes to future, larger event samples from either detector.
The candidate-list search yields no significant excess for any of the 57 objects and provides per-object 90% C.L. upper limits on the flux normalization for $E^{-2}$ and $E^{-2.5}$ spectra.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

If the 2-degree Galactic Centre excess is a genuine neutrino signal rather than a background fluctuation, its fitted parameters ($\hat{n}_s=20.3$, $\hat{\gamma}=3.0$) point to a population of unresolved cosmic-ray accelerators spread over a degree-scale region, a hypothesis a larger deep-sea telescope with Southern-Sky coverage could confirm or reject within a few years.
The distance between the RXJ 1713.7-3946 limits and the two assumed model fluxes (factors of 10.7 and 9.7) implies that roughly an order-of-magnitude gain in sensitivity is needed to test those models; such a gain is plausible with a several-times-larger detector effective volume.
The same likelihood machinery could be repurposed as a live monitoring tool: recomputing the local p-value of the Galactic Centre hotspot after each new data release would show whether the excess grows or decays, providing an early, trial-free test of its reality.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Colleague,

The useful thing to know: this is a conference proceedings from ICRC2019 reporting the first combined ANTARES+IceCube point-like and extended source search in the Southern Sky. The headline is a null result with a factor ~2 sensitivity gain compared to individual searches. That gain is plausible: ANTARES has better angular resolution and Southern-Sky visibility below 100 TeV, IceCube adds statistics, and a joint likelihood over the seven samples is the natural way to exploit both. The paper reuses the published event samples and PDFs from [3] and [4], which is the right move for a first combination and makes the analysis easy to trust. I see no invented entities; the free parameters are standard.

What the paper does well: it is honest about post-trial p-values. The full-Sky point-like search has most significant cluster at post-trial p=0.18, the Galactic Centre extended 2.0-degree search has p=0.03, and no search reaches anything close to 5 sigma. It gives declination-dependent upper limits for E^-2 and E^-2.5 spectra and a dedicated RXJ 1713.7-3946 search. For a proceedings, that is a solid deliverable.

Soft spots: it is a proceedings, so systematic uncertainties are not quantified and several figures say PRELIMINARY. The larger presentation gap is Eq. (3.1). It writes L(n_s, gamma) but the likelihood only shows n_j^s per sample, and the text never states how n_j^s is derived from a common source flux. If the seven n_j^s were literally independent, the test statistic would have seven amplitude parameters and would not be the statistic for a single astrophysical source. I do not think this is load-bearing for the null result—the flux upper limits in Figure 2 show they must have a common normalization with per-sample exposures, and the pseudo-experiments calibrate whatever statistic they actually use—but it is exactly the kind of missing equation a referee would ask to see. The 3% Galactic Centre hotspot is interesting but not significant, and the paper does not overclaim it.

Who this is for: neutrino astrophysicists tracking Southern-sky source constraints and anyone planning joint analyses for KM3NeT or IceCube-Gen2. It deserves a serious referee if the collaboration submits the full analysis to a journal; for the ICRC proceedings, treat it as a useful status report rather than a definitive reference.

Referee Report

2 major / 5 minor

Summary. This proceedings paper reports a combined search for point-like and extended neutrino sources in the Southern Sky, using track and shower events from the nine-year ANTARES point-source analysis and through-going tracks from the seven-year IceCube point-source search. A seven-sample unbinned likelihood (Eq. 3.1) is used to scan the Southern Sky, to test a list of 57 candidate sources, and to perform dedicated searches of the Galactic Centre region, Sagittarius A*, and RX J1713.7-3946. The paper reports no significant excess; the most significant results are a point-like cluster with post-trial p = 0.18 in the full-sky scan and an extended (2°) cluster with post-trial p = 0.03 in the Galactic Centre region. The abstract and conclusions claim an improvement in Southern-Sky sensitivity by a factor of about 2 compared to individual analyses, and 90% confidence level upper limits are presented.

Significance. The combination of ANTARES and IceCube is an important step for Southern-Sky neutrino astronomy, and the claimed factor of about 2 sensitivity gain is an encouraging result. The paper adheres to established likelihood methods, uses signal and background probability density functions from prior peer-reviewed analyses, and applies a Monte-Carlo based trial correction, which are strengths. However, the underspecified coupling between per-sample signal counts and a common source flux means that the statistical hypothesis tested by the quoted p-values and the basis for the sensitivity-gain claim must be clarified before the results can be accepted at face value.

major comments (2)

[Section 3, Eq. (3.1)] The likelihood is written in terms of n_j^s, the signal count in sample j, but the text introduces n_s as the unknown total number of signal events without specifying any relationship such as n_j^s = n_s A_j(gamma). As written, the alternative hypothesis in the test statistic Q (Eq. 3.2) has seven independent signal amplitudes, one per sample, constrained only by the sum n_s = sum_j n_j^s. The reported p-values (e.g., 0.18 in Section 4.1, 0.03 in Section 4.3) and the sensitivity gain are then not p-values and sensitivities for the hypothesis of a single physical source with one flux normalization. The authors should either provide the missing coupling (for instance via sample-dependent acceptance fractions) or explicitly state that the n_j^s are independent; in the latter case the physical interpretation of the limits changes and should be discussed.
[Section 3.1, text after Eq. (3.1)] The statement that the values of the number of signal events ns (ns >= 0.001) and the signal spectral index gamma are optimised is inconsistent with the likelihood, which contains only n_j^s and no explicit n_s. If n_s is a derived quantity, the optimisation should be described in terms of the n_j^s with the constraint n_s = sum_j n_j^s; if n_s is the primary parameter, Eq. (3.1) must be rewritten to show its relation to n_j^s. The present text makes it impossible to reproduce the fitted best-fit values listed in Tables 2, 4, and 5.

minor comments (5)

[Abstract] The phrase 'events from the direction of the Southern Sky' should read 'events with directions in the Southern Sky' to avoid implying a single direction.
[Section 4.1] The phrase 'the location of the fitted cluster being left free to vary within these boundaries' is unclear; if the scan is performed in 1-degree steps, the text should state whether the cluster position is optimized over a continuous range or within each grid cell.
[Section 4.4] There is a missing space in 'Table5'; it should read 'Table 5'.
[Table 3] The notation 'Phi90%E-2.0_nu' in the table header is difficult to parse; consider a clearer formatting such as 'Phi^{90%}_{nu,E^{-2.0}}'.
[Section 4.3] The best-fit position for the 2.0-degree extended search, (274.1 degrees, -40.1 degrees), is about 13 degrees away from the Galactic Centre; the text should explicitly note that the excess is not centered on the Galactic Centre, which is relevant for the interpretation of the 'Galactic Centre region' search.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the combined-search results are statistical outputs from external data samples and published PDFs, with no fitted quantity renamed as a prediction.

full rationale

The paper reports an experimental search: it takes data samples and signal/background PDFs from the published ANTARES [3] and IceCube [4] point-source analyses, constructs the joint likelihood in Eq. (3.1), maximizes it over the signal normalization and spectral index, and compares the resulting test statistic to background-only pseudo-experiments. The quoted post-trial p-values and 90% C.L. upper limits are observed outcomes of that procedure, not quantities derived from themselves. The sensitivity gain of about a factor of 2 is an estimate from pseudo-experiments and from the combined likelihood, not a claim that a fitted parameter has been independently predicted. Citations to the collaborations' own earlier papers provide the concrete data samples and PDF parameterizations, which are external inputs with stated assumptions; they are not used to assert the conclusion of this paper. The possible ambiguity in Eq. (3.1) about how the per-sample signal counts n_j^s relate to a common source flux is a modeling/presentation concern, but it does not make the analysis circular. Hence the correct finding is no significant circularity, score 0.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The central claim is a null result, so the main unverified inputs are inherited PDFs from earlier analyses and the background-only pseudo-experiment simulation. No new particles, forces, dimensions, or conserved quantities are introduced, and no free parameters beyond the fitted signal parameters and scanned source positions appear in the likelihood.

free parameters (3)

signal event count n_s = 5.7 (full-sky best, point-like); 20.3 (Galactic Centre, 2.0 deg); per-candidate values in Table 3
Optimized in the likelihood of Eq. (3.1) for each source hypothesis; it directly determines the test statistic Q and the quoted p-values.
spectral index gamma = 2.5 (full-sky best, point-like); 3.0 (Galactic Centre, 2.0 deg); per-candidate values in Table 3
Fitted jointly with n_s within the range 1 to 4; it enters the assumed E^-gamma signal spectrum.
source position (RA, Dec) = (213.2 deg, -40.8 deg) for full-sky best; (274.1 deg, -40.1 deg) for Galactic Centre search
Left free within 1 degree by 1 degree cells in the scans; the trial factor in the pseudo-experiments accounts for scanning many positions.

assumptions (5)

standard math The maximum likelihood defined in Eq. (3.1) over the product of per-sample and per-event terms is a valid inference framework.
Used without proof as the statistical basis for the search.
domain assumption The signal and background PDFs from the nine-year ANTARES analysis [3] and the seven-year IceCube analysis [4] remain valid when the samples are combined.
Section 3 states the PDF definitions differ slightly for ANTARES and IceCube and can be found in [3] and [4]; no new validation is presented.
domain assumption Time-randomized pseudo-experiments produce background-only test statistic distributions that match the real atmospheric background.
Section 3, last paragraph: the post-trial p-values are computed by comparing observed Q to this pseudo-experiment distribution.
domain assumption Extended sources are modeled as Gaussian with effective width sigma_eff = sqrt(sigma^2 + sigma_s^2) for IceCube and as a MC PSF convolved with a Gaussian for ANTARES.
Section 3, extended-source paragraph; this shape assumption drives the extended-source limits.
domain assumption The RXJ 1713.7-3946 neutrino spectra follow the two published models [7,8] with parameters in Table 6.
Section 4.5 and Eq. (4.1); the upper limit ratios depend directly on these assumed spectra.

how reviews work

0 comments

Cite this review

Pith. "Pith review of ANTARES and IceCube combined search for neutrino point-like and extended sources in the Southern Sky." pith.science (2026). https://pith.science/paper/FCNAADVH

@misc{pith2026190807439,
  author       = {Pith},
  title        = {Pith review of: ANTARES and IceCube combined search for neutrino point-like and extended sources in the Southern Sky},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FCNAADVH}},
  note         = {Machine review of arXiv:1908.07439}
}

abstract

The ANTARES neutrino telescope, located in the Mediterranean Sea, and the IceCube neutrino observatory, located at the geographic South Pole, both search for cosmic neutrino events with an instantaneous full-sky field of view. The different characteristics of the two telescopes, in particular the larger instrumented volume of IceCube and the better visibility towards the Southern Sky for neutrino energies below 100 TeV of ANTARES, are exploited in a combined search for point-like and extended sources. The sensitivity to neutrino sources located in the Southern Sky is improved by a factor of $\sim$2 compared to individual studies. The data samples used in this analysis correspond to all track-like and shower-like events from the direction of the Southern Sky which were included in the nine-year ANTARES point-source analysis, combined with the through-going track-like events used in the seven-year IceCube point-source search. In this analysis, the Southern Sky is scanned for possible excesses of neutrinos that might indicate the presence of a source, while the coordinates of predefined candidate neutrinos sources are also evaluated in order to limit the penalty of trials. In addition, special focus is given to the region around the Galactic Centre, treated as an extended neutrinos source, and to the location of the supernova remnant RXJ 1713.7-3946. The result of this combined search for galactic and extra-galactic neutrino sources in the Southern Sky is reported here. No significant evidence of cosmic neutrino sources is found and flux upper limits from the various searches are presented.

Figures

Figures reproduced from arXiv: 1908.07439 by the authors.

**Figure 1.** Sky map in equatorial coordinates of the pre-trial p-values obtained in the Southern-sky search for the point-like source hypothesis. The red contour indicates the location of the most significant cluster. sinδ −1 −0.8 −0.6 −0.4 −0.2 0 ] -1 s -2 [GeV cm ν /dE Φ d 2 ν E −9 10 −8 10 −7 10 90% C.L. Sensitivity and Limits for γ = 2.0 IceCube 7 years ANTARES 9 years IceCube+ANTARES IceCube+ANTARES Limits 90% C.L. Sensiti… view at source ↗

**Figure 2.** Upper limits at 90% C.L. on the signal flux from the analysed candidates (green dots). An unbroken E −γ ν neutrino spectrum is assumed, with γ = 2.0 (left) and γ = 2.5 (right). The green line indicates the sensitivity of the combined analysis. The dashed curves indicate the sensitivities for the IceCube (blue) and ANTARES (red) individual analyses. 4.3 Galactic Centre region The third search is similar to the Southe… view at source ↗

**Figure 3.** Left: 90% C.L. limits on the neutrino flux of the Galactic Centre region search assuming an E −2.0 ν spectrum for different source extensions σs . Right: discovery flux (solid red), median sensitivity (solid blue) and 90% C.L. upper limits (solid green) for the search at the location of Sagittarius A* assuming an E −2.0 ν spectrum as a function of the angular extension σs . The results for the point-like source assu… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

12 extracted references · 7 canonical work pages

[3]

Albert et al., Phys

ANTARES Collaboration, A. Albert et al., Phys. Rev. D96 (2017) 082001

2017
[4]

IceCube Collaboration, M. G. Aartsen et al., Astrophys. J. 835 (2017) 151

2017
[1]

Ageron et al., Nucl

ANTARES Collaboration, M. Ageron et al., Nucl. Instrum. Meth. A656 (2011) 11

2011
[2]

IceCube Collaboration, M. G. Aartsen et al., JINST 12 (2017) P03012

2017
[5]

Collaboration, F

H.E.S.S. Collaboration, F. Aharonian, Astron. Astrophys. 467 (2007) 1075–1080

work page 2007
[6]

Neyman, Phil

J. Neyman, Phil. Trans. Roy. Soc. Lond. A236 (1937) 333

1937
[7]

Kappes, J

A. Kappes, J. Hinton, C. Stegmann, and F. A. Aharonian, Astrophys. J. 656 (2007) 870–896. [Erratum: Astrophys. J. 661 (2007) 1348]

work page 2007
[8]

Trovato, R

KM3NeT Collaboration, A. Trovato, R. Coniglione, P. Sapienza, and J. Barrios-Martí, PoS(ICRC2017)999 (2018)

work page 2018

Show all 12 references

[9]

Vissani, Astropart

F. Vissani, Astropart. Phys. 26 (2006) 310–313

2006
[10]

F. L. Villante and F. Vissani, Phys. Rev. D78 (2008) 103007

2008
[11]

Vissani and F

F. Vissani and F. L. Villante, Nucl. Instrum. Meth. A588 (2008) 123–129

2008
[12]

Abdalla et al., Astron

HESS Collaboration, H. Abdalla et al., Astron. Astrophys. 612 (2018) A6. 8

2018

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [3]

Albert et al., Phys

ANTARES Collaboration, A. Albert et al., Phys. Rev. D96 (2017) 082001

2017

[2] [4]

IceCube Collaboration, M. G. Aartsen et al., Astrophys. J. 835 (2017) 151

2017

[3] [1]

Ageron et al., Nucl

ANTARES Collaboration, M. Ageron et al., Nucl. Instrum. Meth. A656 (2011) 11

2011

[4] [2]

IceCube Collaboration, M. G. Aartsen et al., JINST 12 (2017) P03012

2017

[5] [5]

Collaboration, F

H.E.S.S. Collaboration, F. Aharonian, Astron. Astrophys. 467 (2007) 1075–1080

work page 2007

[6] [6]

Neyman, Phil

J. Neyman, Phil. Trans. Roy. Soc. Lond. A236 (1937) 333

1937

[7] [7]

Kappes, J

A. Kappes, J. Hinton, C. Stegmann, and F. A. Aharonian, Astrophys. J. 656 (2007) 870–896. [Erratum: Astrophys. J. 661 (2007) 1348]

work page 2007

[8] [8]

Trovato, R

KM3NeT Collaboration, A. Trovato, R. Coniglione, P. Sapienza, and J. Barrios-Martí, PoS(ICRC2017)999 (2018)

work page 2018