Pith. sign in

REVIEW 4 major objections 5 minor 21 references

Low Energy Cosmic Ray Spectrum from 250 TeV to 10 PeV using IceTop

T0 review · 4 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 Using a new two-station infill trigger, IceTop measures the all-particle cosmic ray energy spectrum from 250 TeV to 10 PeV, and both hadronic interaction models yield a bend near the knee.

desk verdict Genuinely new IceTop low-energy spectrum from a new infill trigger, but the bend claim rests on composition systematics not propagated through the response matrix and on visual inspection. read the letter →

arxiv 1908.07143 v1 pith:KZ2G4IUS submitted 2019-08-20 astro-ph.HE

Ramesh Koirala , Thomas K. Gaisser (for the IceCube Collaboration) This is my paper · ORCID
classification astro-ph.HE
keywords cosmicrayenergyspectrumIceTopkneeall-particleairshowerinfilltriggerBayesianunfolding250TeVto10Pe
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 reports a new measurement of the all-particle cosmic ray energy spectrum from 250 TeV to 10 PeV using the IceTop surface array. The authors implemented a two-station trigger on the dense infill array, which lowers IceTop's energy threshold by almost an order of magnitude and narrows the gap to direct balloon and satellite measurements. Using random-forest reconstruction and iterative Bayesian unfolding, they obtain a spectrum for each of two hadronic interaction models; both show a bend in the knee region around a few PeV. If correct, the result connects low-energy direct measurements with the higher-energy IceTop spectrum and provides an independent cross-check for ground-based arrays in the overlapping energy range.

What carries the argument

The enabling object is the two-station infill trigger: six infill stations closer than 50 m apart, arranged as four pairs, register an event when any pair is hit within 200 ns, and the entire array's local coincidences within a 10 microsecond window are read out. This trigger collects small air showers that would otherwise be lost. A random forest regressor trained on Monte Carlo predicts core position, zenith angle, and energy from tank charges and hit times; an iterative Bayesian unfolding converts the reconstructed energy distribution into a true energy distribution using a response matrix. The effective area, computed from the Monte Carlo with a modified H3a composition model and Sibyll 2.1, converts the unfolded event count into a flux.

What would settle it

Measure the mean logarithmic mass of cosmic rays in the 100 TeV to 1 PeV range with an independent technique, such as a dedicated air-shower array or a future direct experiment. If that measurement falls outside the envelope spanned by the modified H3a, GST, GSF, and modified Polygonato models, the response matrix used here is biased, and the reconstructed energy spectrum—and hence the bend—would shift by more than the reported systematic uncertainty. Alternatively, if a future hadronic interaction model changes the predicted detector response by more than the quoted systematic band, the bend could move or disappear.

Watch

Extended reading notes

Core claim

The central claim is that IceTop can measure the all-particle cosmic ray flux from 250 TeV to 10 PeV, and that the resulting spectrum exhibits a bend near the knee of the cosmic ray spectrum. The authors lower the energy threshold by triggering on events that hit at least one of four closely spaced infill station pairs, collect 7.4 million events in 330 days of livetime, and reconstruct each shower's core, direction, and energy with a random forest trained on air-shower Monte Carlo. After quality cuts and iterative Bayesian unfolding with a response matrix built from simulation, two spectra are derived, one with Sibyll 2.1 and one with QGSJetII-04. Both spectra show a bend in the knee region, are compatible within systematics with the previous 3-year IceTop spectrum in the overlap region, and overlap within systematic errors with an independent ground-based measurement near 300 TeV.

Load-bearing premise

The load-bearing premise is that the true cosmic-ray composition over 250 TeV to 10 PeV lies within the range spanned by the four composition models used in the simulation; if it does not, the reconstructed energies and the bend would shift by more than the quoted uncertainty.

Editorial extensions

If this is right

  • The IceTop energy threshold drops from about 1.5 PeV to 250 TeV, so IceTop now spans a range previously covered only by other ground arrays.
  • The spectrum provides an overlap region with direct measurements, allowing a consistency check of the transition from direct to indirect techniques.
  • The bend near the knee is visible in both hadronic interaction models, suggesting the feature is not an artifact of one particular interaction model.
  • The new low-energy events add an independent data set for composition studies in the 100 TeV to 1 PeV range.
  • The comparison with the higher-energy IceTop spectrum, within 8.5%, indicates the two analyses are consistent despite different triggers, reconstruction, and energy scales.

Reading between the lines

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

  • The two-station trigger technique could be pushed further: a denser infill or a lower local-coincidence threshold might extend IceTop's reach down toward 100 TeV, almost completely closing the gap to direct measurements.
  • The composition-model systematics imply that an independent measurement of the mean logarithmic mass in the 100 TeV to 1 PeV range would directly test the reliability of the reconstructed energies; if the true composition lies outside the spanned models, the bend's location could shift.
  • If the bend is confirmed at the same energy by future arrays, it strengthens the case that the knee is a source or propagation effect rather than a detector artifact.
  • The unfolding prior and stopping-criterion choice could affect the tilt of the spectrum; a public release of the unfolded counts would let other groups test different priors.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The paper reports a measurement of the all-particle cosmic-ray energy spectrum from 250 TeV to 10 PeV using IceTop, based on a new two-station trigger that makes use of four pairs of densely spaced infill stations. The analysis uses roughly one year of data (May 2016 through April 2017, 330.43 days livetime), reconstructs core position, zenith angle, and energy with random forest regressors trained on Monte Carlo simulations, and applies an iterative Bayesian unfolding with a spline regularization. Fluxes are computed for two hadronic interaction models, Sibyll 2.1 and QGSJetII-04, and both are reported to show a bend near the knee. The new spectrum is compared with HAWC, KASCADE, Tunka, Tibet III, and ATIC-02, and is stated to overlap with HAWC near 300 TeV and to be compatible with the higher-energy IceTop spectrum within systematics.

Significance. If the measurement is correct, it is valuable because it narrows the gap between direct cosmic-ray measurements (up to ~100 TeV) and the previous IceTop surface-array measurements (above ~1.5 PeV), and it provides an independent cross-check of HAWC and other air-shower experiments in a region where composition and hadronic-interaction uncertainties are large. The paper has clear strengths: a new trigger and filter that extend the detector acceptance, a Monte Carlo based reconstruction with documented resolutions, an explicit accounting of several systematic uncertainties, and a repeat of the full analysis with two hadronic interaction models. The main limitation is that the central claim of a knee-region bend is not yet quantitatively supported: the paper provides no closure test for the unfolding, no statistical significance for the bend, and no flux table, and the composition systematic is treated in a way that does not propagate composition changes through the energy reconstruction.

major comments (4)
  1. [Sec. 3.1] The composition systematic is computed by evaluating the flux for alternate composition models 'using the same response matrix'. This does not propagate composition differences through the random forest energy reconstruction or through the unfolding response matrix itself. A different primary mix changes the mapping from tank charges and hit patterns to primary energy, so a response matrix fixed to the modified H3a model can bias reconstructed energies even if alternate flux weights are later applied. Because the spectrum is steep, a few-percent energy-scale shift that grows with energy can alter or even produce the reported bend. Please retrain or re-weight the full chain for at least one alternate composition model, or provide a demonstration that the energy estimator is composition-insensitive in this energy range.
  2. [Sec. 3, Sec. 4] There is no closure test showing that the reconstruction and unfolding chain recovers an injected Monte Carlo spectrum within statistical and systematic uncertainties. Given that the reported bend is a spectral feature across a narrow energy range, a closure test is needed to show that the unfolding plus quality cuts do not bias the spectral shape, especially at the lowest and highest energies of the reported range. Please add a closure test, ideally including a spectrum with a sharp knee, and report the pulled residuals per bin.
  3. [Sec. 5, Fig. 4] The conclusion that 'both measured energy spectra show a bend around the knee region' is based on visual inspection of scaled plots. No fit with and without a break is shown, and no statistical or systematic significance for the bend is quoted. Please fit the data with a power law and a broken power law, and report the improvement in fit quality and the best-fit break position and its uncertainty from the combined statistical and systematic errors.
  4. [Sec. 4] The paper does not provide a numerical flux table. For a measurement that is intended to fill the gap between direct and indirect experiments and to cross-check HAWC and others, the binned fluxes, bin centers, statistical errors, and systematic errors must be tabulated (or made available in a repository cited in the paper). Without this, the central result cannot be used or scrutinized by the community.
minor comments (5)
  1. [Fig. 1 caption] The caption contains a typo: 'Sybill 2.1' should be 'Sibyll 2.1'.
  2. [Ref. [13]] Reference [13] lists the author as 'N. Smrinov'; the correct spelling is 'N. Smirnov'.
  3. [Eq. (4.1)] In Equation (4.1), the notation '∆lnEπ' is ambiguous; it should be written as Δ ln E · π or with explicit parentheses to avoid the appearance that π multiplies the logarithm.
  4. [Fig. 5] The caption states that the shaded region indicates systematic uncertainties but does not specify whether it is the uncertainty of this analysis only or includes the compared data sets. Please clarify.
  5. [Sec. 3.1] The text describing the atmospheric systematic would be clearer if the pressure correction factor were quoted with its uncertainty rather than only 'around 5%'.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the new IceTop spectrum is derived from data through MC-trained reconstruction and unfolding, with external cross-checks and no fitted parameter renamed as a prediction.

full rationale

The paper's central claim is an experimentally measured all-particle cosmic ray spectrum from 250 TeV to 10 PeV using a new two-station infill trigger. The derivation chain is: (1) collect events with a new hardware trigger, (2) reconstruct core, zenith, and energy with random forests trained on air-shower Monte Carlo, (3) apply quality cuts, (4) unfold the reconstructed energy distribution with an iterative Bayesian method using a simulated response matrix, and (5) divide by livetime, solid angle, and effective area to obtain flux. No parameter is fitted to the final binned spectrum and then presented as a prediction; the unfolded distribution is an output from the data, not an input. The appearance of a bend in the knee region is reported for both Sibyll 2.1 and QGSJetII-04 analyses, which reduces the chance that the feature is an artifact of one hadronic model. The paper also compares the result with HAWC, KASCADE, Tunka, Tibet III, and ATIC-02, providing external checks. The composition systematic is estimated by reweighting alternate composition models through the same response matrix; this is a recognized limitation of the systematic treatment rather than a circular derivation of the spectrum, because the reconstructed data distribution is not itself set by the composition prior. Self-citations to earlier IceTop spectrum analyses are used as context and comparison, not as a load-bearing uniqueness argument. Therefore no circular step meeting the evidentiary standard is present.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The paper is an experimental measurement with no new theoretical entities. The main external inputs are the hadronic interaction models, composition models, and detector simulation. The largest load-bearing assumption is the composition model, which the paper itself identifies as the dominant systematic.

free parameters (3)
  • Effective area sigmoid fit parameters = not specified
    The effective area shown in Fig. 1 (right) is fitted with an energy-dependent sigmoid function (Section 3.1), and the flux calculation in Eq. 4.1 uses this fitted Aeff. The systematic uncertainty is derived from the errors on these parameters.
  • Unfolding spline regularization strength = not specified
    Section 3 states that a spline fit is performed on the unfolded energy spectrum after each iteration to regularize the updated prior, but the smoothing strength is not given. The unfolded spectrum depends on this choice.
  • Quality cut thresholds = largest charge <= 75% total; sum of two largest <= 90%; >= 42 stations; total charge > 103.8 VEM excluded
    These cuts (Section 3) are chosen by hand to remove poorly reconstructed events. They affect the event sample and thus the final flux.
assumptions (4)
  • domain assumption Air shower simulations with Sibyll 2.1 and QGSJetII-04 accurately model the development of cosmic-ray air showers in the energy range 250 TeV to 10 PeV.
    The entire reconstruction, effective area, and response matrix are based on these simulations (Sections 3 and 3.1).
  • domain assumption The modified H3a composition model, and the alternate models GST, GSF, and modified Polygonato, bracket the true cosmic-ray composition.
    The composition systematic is estimated as the maximum deviation between fluxes computed with these models (Section 3.1). If the true composition is outside this bracket, the energy scale is biased.
  • domain assumption The Monte Carlo detector simulation accurately reproduces IceTop trigger, charge calibration, snow conditions, and atmospheric profiles.
    The response matrix and effective area are taken from simulation; deviations are only partially corrected (pressure correction, snow height estimate). Section 3.
  • domain assumption The random forest regression trained on simulated events generalizes to experimental data.
    The reconstruction models for core, zenith, and energy are trained and evaluated only on MC (Section 3, Fig. 2).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Low Energy Cosmic Ray Spectrum from 250 TeV to 10 PeV using IceTop." pith.science (2026). https://pith.science/paper/KZ2G4IUS

@misc{pith2026190807143,
  author       = {Pith},
  title        = {Pith review of: Low Energy Cosmic Ray Spectrum from 250 TeV to 10 PeV using IceTop},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KZ2G4IUS}},
  note         = {Machine review of arXiv:1908.07143}
}
read the original abstract

Using IceTop, the surface component of the IceCube Neutrino Observatory, the all-particle cosmic ray energy spectrum has been determined above a few PeV. The measured energy spectrum leaves a gap of more than a decade in energy to direct measurements by balloon and satellite experiments. In this analysis, we lowered the energy threshold of IceTop to 250 TeV, narrowing the gap between IceTop and direct measurements. In order to collect lower energy events, we implemented a new trigger that uses four pairs of infill stations for which the separation between stations is less than 50 m, compared to 125 m for the main array. The new trigger collects data from the entire array for events with hits on at least one infill pair. The low-energy extension of the all-particle cosmic ray energy spectrum using these IceTop events is measured and is compared with the energy spectrum from HAWC and other experiments. Air shower simulations with two different hadronic interaction models, Sibyll 2.1 and QGSJetII-04, are used in this analysis and an energy spectrum for each model is produced. Both measured energy spectra show a bend around the knee region.

Figures

Figures reproduced from arXiv: 1908.07143 by the authors.

Figure 1
Figure 1. Left: IceTop geometry with positions of all tanks. Nearby infill boundary includes 6 out of 8 infill stations. Tanks inside this boundary at the center are used to define a two-station trigger. Right: Effective area calculated using MC generated with modified H3a composition model, Sybill 2.1 hadronic interaction model, and cosine of zenith angle greater than or equal to 0.9. The line is a fit through the simulated … view at source ↗
Figure 2
Figure 2. Left: Core resolution in meter; Middle: zenith resolution in degree; Right: energy resolution in unit-less quantity. energy reconstruction are the charge deposited on tanks, distance of hit tanks from the shower core, and the zenith angle of an air shower. Charge is calibrated in units of vertical equivalent muons (VEM) [3] [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. The individual systematic uncertainties for each energy bins. Total systematic uncertainty is the sum of individual uncertainties added in quadrature. 3.1 Systematic Uncertainties Systematic uncertainties are calculated by keeping all conditions constant except the feature under investigation. The systematics uncertainties due to the hadronic interaction models is consid￾ered separately. The major systematic uncerta… view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Left: The all-particle cosmic ray energy spectrum using IceTop 2016 data. The analysis is done using simulations with Sibyll 2.1 as the hadronic interaction model. Right: The all-particle cosmic ray energy spectra using simulations with Sibyll2.1 and QGSJetII-04 as had…
Figure 5
Figure 5. Figure 5: Cosmic ray flux using IceTop 2016 data scaled by E 1.65 and compared with flux from other experiments. This analysis and HAWC’s energy spectrum analysis use different hadronic interaction models. The shaded region indicates the systematic uncertainties. The all-particl…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

21 extracted references · 16 canonical work pages

  1. [1]

    IceCube Collaboration, M. G. Aartsen et al., Phys. Rev. D88 (2013) 042004

    2013

  2. [2]

    IceCube Collaboration, M. G. Aartsen et al., arXiv:1906.04317

    arXiv 1906

  3. [3]

    Abbasi et al., NIM 700 (2013) 188–220

    IceCube Collaboration, R. Abbasi et al., NIM 700 (2013) 188–220

    work page 2013

  4. [4]

    Abbasi et al., Astropart

    IceCube Collaboration, R. Abbasi et al., Astropart. Phys. 44 (2013) 40–58

    work page 2013

  5. [5]

    Abbasi et al., Astroparticle Physics 42 (Feb, 2013) 15–32

    IceCube Collaboration, R. Abbasi et al., Astroparticle Physics 42 (Feb, 2013) 15–32

    work page 2013

  6. [6]

    Rawlins, J

    IceCube Collaboration, K. Rawlins, J. Phys. Conf. Ser .718 (2016) 052033

    2016

  7. [7]

    Breiman, Machine Learning 45 (2001) 5–32

    L. Breiman, Machine Learning 45 (2001) 5–32

    work page 2001

  8. [8]

    James et al., An Introduction to Statistical Learning

    G. James et al., An Introduction to Statistical Learning . Springer-Verlag New York, 2013

    work page 2013

Show all 21 references
  1. [9]

    D’Agostini, Nucl

    G. D’Agostini, Nucl. Instrum. Meth. A362 (1995) 487–498

    1995

  2. [10]

    D’Agostini, arXiv:1010.0632

    G. D’Agostini, arXiv:1010.0632

    arXiv

  3. [11]

    Bourbeau and Z

    J. Bourbeau and Z. Hampel-Arias, arXiv:1806.03350

    arXiv

  4. [12]

    A. N. Kolmogorov, Giornale dell’Istituto Italiano degli Attuari 4 (1933) 83–91

    1933

  5. [13]

    Smrinov, Ann

    N. Smrinov, Ann. Math. Statistics 19 (1948) 279–281

    1948

  6. [14]

    Jeffreys, Proc

    H. Jeffreys, Proc. Roy. Soc. A 186 (1945)

    1945

  7. [15]

    T. K. Gaisser, T. Stanev, and S. Tilav, Front. Phys.(Beijing) 8 (2013) 748–758

    2013

  8. [16]

    Dembinski et al., PoS(ICRC2017)533 (2018)

    H. Dembinski et al., PoS(ICRC2017)533 (2018)

    2018

  9. [17]

    Alfaro et al., Phys

    HA WCCollaboration, R. Alfaro et al., Phys. Rev. D96 (2017) 122001

    2017

  10. [18]

    Antoni et al., Astropart

    KASCADE Collaboration, T. Antoni et al., Astropart. Phys. 24 (2005) 1–25

    2005

  11. [19]

    V . V . Prosin et al.,NIM 756 (2014) 94 – 101

    2014

  12. [20]

    Amenomori et al., Astrophys

    TIBET III Collaboration, M. Amenomori et al., Astrophys. J. 678 (2008) 1165–1179

    2008

  13. [21]

    A. D. Panov et al., Bull. Russ. Acad. Sci. Phys 73 (Jun, 2009) 564–567. 8

    2009

Pith tools

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