QGSJET-II: towards reliable description of very high energy hadronic interactions

The thing to know: this is the announcement paper for QGSJET-II, and the substantive advance is genuine. Prior QGSJET kept at most a single triple-Pomeron vertex; here the enhanced diagrams are resummed into a closed total eikonal (Fig. 6) for the uncut amplitude and for the relevant unitarity cuts, and the result is folded into an event-by-event Monte Carlo rather than applied as an average correction. That is the piece the field has been waiting for, and doing it for cut topologies — not just the elastic amplitude — is what makes it usable in a shower generator. The paper is also candid about being a summary, which I appreciate even though it costs it on completeness. The full-versus-dashed curves in Figs. 7 and 8 are a real internal ablation, and getting the pp total cross section and F2 at Q0² = 2 GeV² out of one parameter set is a nontrivial consistency check, postdiction or not. I would not lean as hard on the calibration-circularity point as the reader does; the ablation is the evidence, and it holds up on its own terms. Where I do think the paper overreaches is the loop sentence in Section 3, just before Fig. 6: 'neglecting subdominant contributions of Pomeron loops.' The stress-test note lands. The resummation is all-orders only within the fan and generalized-fan topologies, and loops are precisely the class whose relative weight grows with energy. No numerical bound is given at 10¹⁹-10²⁰ eV, so 'reliable model extrapolation into ultra-high energy domain' in the abstract is asserted, not demonstrated. The soft-dominance assumption for Pomeron-Pomeron vertices is at least flagged as an assumption; the loop truncation is not flagged at all, and the ablation cannot test it. Secondary: the derivation and sampling algorithm sit in ref. [15], 'in preparation', and the nuclear/EAS consequences in [16], 'these proceedings', so nothing here is independently checkable. For a four-page proceedings that is normal, but it means the paper's claims rest on work I cannot see. Who should read it: cosmic-ray people who use QGSJET and need to know what changed, and anyone tracking screening corrections in hadronic generators. Worth a serious referee, with the explicit ask that the companion paper quantify the neglected loops and give parameter values and fit quality.

The central claim (abstract:s3, abstract:s2) is that enhanced Pomeron diagrams are resummed to all orders into a closed total eikonal (Fig. 6), and that this is what makes the model trustworthy at ultra-high energy. The load-bearing sentence sits in Section 3, immediately before Fig. 6: 'Then, neglecting subdominant contributions of Pomeron loops, the final result for hadron-hadron interaction eikonal can be represented in the graphical form as shown in Fig. 6.' So the resummation is all-orders only within the loopless subset (fan and generalized-fan topologies of Figs. 4 and 5). Pomeron loops are exactly the class of enhanced diagrams whose relative weight is expected to grow with energy; the fan series alone is the standard example of a set that does not by itself preserve t-channel unitarity asymptotically. The paper gives no estimate of the loop contribution at, say, 10^19-10^20 eV, so 'reliable model extrapolation into ultra-high energy domain' (abstract:s2) is asserted rather than bounded. This is a sibling of the reader's soft-dominance point, not the same point: soft dominance fixes which Pomerons may enter a multi-Pomeron vertex; loop neglect fixes which topologies are summed. Both are required for Fig. 6 to close, but only the first is flagged. The internal ablation in Figs. 7 and 8 (full versus dashed 'without enhanced diagrams') tests enhanced-versus-none, not loops-versus-no-loops, so it cannot speak to this. The derivation itself is deferred to reference [15], listed as 'in preparation', so the size of the neglected loops cannot be checked from this document either.

The reader's verdict of CONDITIONAL is right, and the reader's rationale already notes that the derivation and sampling algorithm live in reference [15], 'in preparation'. Where I would sharpen rather than redirect: the reader's weakest_assumption is the soft-dominance premise for Pomeron-Pomeron vertices (Section 3, Fig. 3). That premise is real and is stated as an assumption, but it is one of two approximations that the Fig. 6 closure depends on. The second, stated one sentence earlier in the same paragraph, is the neglect of Pomeron loops. That one is more directly in tension with the paper's own framing, because the abstract sells the result as an all-order resummation valid for ultra-high-energy extrapolation, and loops are the diagrams whose relative importance is conventionally expected to rise with energy. The paper does not quantify the neglected terms anywhere, and the only ablation it shows (full versus dashed curves in Figs. 7 and 8) switches enhanced diagrams on and off wholesale, so it cannot isolate the loop contribution. I therefore read the soft-dominance assumption as necessary but not sufficient for the central claim, and I would put the loop-neglect sentence at the top of the list of things a referee should ask to see bounded. On the evidence side, I share the reader's observation that Figs. 7 and 8 are shown after a re-calibration on a wider accelerator data set and that both observables flow from the same soft and semihard Pomeron parameters, so 'consistent description of cross sections and structure functions' is closer to a two-observable postdiction than to two independent tests. That is a correctness-risk point about how strongly the plots support the claim, not a claim that the numbers are wrong. The paper is honest about being a summary, and the graphical recursions plus the internal ablation are genuine content; nothing here reads as unsound, only as under-quantified. The verdict stays CONDITIONAL, and the concrete test above is the single check that would most cleanly separate 'defensible approximation' from 'unbounded extrapolation claim'.

If the resummation holds, this is a substantive step for the field: it replaces the linear (independent-cascade) Pomeron picture that underlies essentially all hadronic event generators used in air-shower work with a scheme in which screening is sampled event-by-event, and it does so without introducing new parameters relative to the existing model. Two features deserve explicit credit. First, the simultaneous description of two very different observables — the total pp cross section and the low-Q^2 proton structure function — with a single parameter set is a non-trivial consistency check, since the same eikonal controls both. Second, the observation in §4 that the semihard eikonal is no longer factorizable while inclusive high-pt jet spectra remain unaffected by AGK cancellations is a sharp, checkable statement: it predicts where the scheme must reproduce QCD factorization and where it must deviate, and it is the kind of claim that can be falsified against collider jet data. The nuclear extension, in which Pomerons of one enhanced graph may couple to different nucleons, is also a physically consequential prediction (violation of superposition for average EAS characteristics). The

This is the first general update of QGSJET, a hadronic interaction model that cosmic-ray groups use to simulate the cascades of particles that air showers are made of. The problem it attacks is stated early: the standard treatment assumes that the individual parton cascades exchanged in a collision, modelled as Pomerons, develop independently, and that assumption fails at high energy and small impact parameter, where many cascades overlap and screen one another. The paper's move is to resum the 'enhanced' diagrams that describe Pomeron-Pomeron interactions to all orders, both for the uncut elastic amplitude and for every unitarity cut corresponding to a physical final state, arriving at a closed expression for the total eikonal. The payoff claimed is consistency rather than a single number: one parameter set reproduces the measured total proton-proton cross section and the proton structure function $F_2(x,Q_0^2)$ at $Q_0^2=2$ GeV$^2$, and the resulting screening is then applied explicitly within sampled events instead of only on average. The reason a reader outside the field should care is that essentially all cosmic-ray data above $10^{14}$ eV are reconstructed through models of this kind, and the ultra-high-energy extrapolation is precisely where the independent-cascade assumption gives way.

What carries the argument. The carrying object is the total eikonal $X_{tot}(s,b) = X_P(s,b) + X_{sh}(s,b)$, the sum of a soft and a semihard Pomeron eikonal, out of which total, inelastic and configuration probabilities are obtained by applying the Abramovskii-Gribov-Kancheli cutting rules to the multi-Pomeron diagrams. The new ingredient is the replacement of each simple Pomeron exchange in that sum by a re-summed 'generalized fan' built from the recursive equations of Figs. 4 and 5. The recursion closes only because of a stated physical assumption about where Pomeron-Pomeron vertices live, and the closure is what turns an infinite class of enhanced graphs into a single computable eikonal that a Monte Carlo can then cut and sample.

Core claim. The central claim is that the non-linear, Pomeron-Pomeron interaction effects which spoil the independent-cascade picture can be re-summed to all orders into a usable, closed form, and that the result is a working model. Concretely, Ostapchenko computes the 'fan' diagram contribution from a recursive graphical equation, extends it to a 'generalized fan' that also includes vertices with fans attached to both the projectile and the target, and then, neglecting subdominant Pomeron-loop contributions, writes the full hadron-hadron interaction eikonal in the graphical form of Fig. 6. The elastic amplitude is the usual multi-Pomeron sum with each elementary exchange replaced by that full set of diagrams. The author then takes all unitarity cuts of those diagrams, re-sums cuts of given types, and builds a Monte Carlo procedure that samples interaction events of complicated topologies while accounting for screening in each particular cut-Pomeron configuration. The evidence offered for the whole construction is that the same parameter set lands on the measured total $pp$ cross section (Fig. 7) and on the proton structure function at $Q_0^2=2$ GeV$^2$ (Fig. 8), with and without the enhanced contributions shown side by side. Hadron-nucleus and nucleus-nucleus collisions are handled the same way with no new parameters, with the added feature that Pomerons of one enhanced graph may couple to different nucleons, which strengthens screening and breaks the superposition picture for nucleus-induced showers. A secondary observation is that the semihard contribution to the eikonal is no longer expressible in the factorized form assumed by earlier hadronic generators, though AGK cancellations keep inclusive high-$p_t$ jet spectra factorizable.

The premise it rests on. The load-bearing premise, stated by the author in Section 3, is that Pomeron-Pomeron interactions are dominated by partonic processes at low virtuality, $|q^2| < Q_0^2$, so that multi-Pomeron vertices involve only soft Pomerons or the soft ends of semihard ones.

Cosmic-ray air-shower simulations above $10^{14}$ eV can be generated with screening corrections applied inside each simulated collision, rather than folded in as an average correction to the interaction probability.

The same parameter set should describe hadron-hadron cross sections and hadron structure functions simultaneously, removing a tuning tension between the two in earlier generators.

For nucleus-induced showers, the superposition of independent nucleon-nucleon collisions is expected to fail even for average shower characteristics, because one enhanced graph can couple to several nucleons.

Inclusive high-$p_t$ jet production should remain factorizable despite the new non-factorizable terms in the eikonal, because the AGK cancellations remove those diagrams from the inclusive jet spectrum.

The model is claimed to extrapolate into the ultra-high-energy domain without the independent-Pomeron assumption, which is the regime where previous hadronic generators had no controlled answer.

Because the closure of the fan recursion rests on Pomeron-Pomeron vertices being soft, the model's screening strength is effectively a one-parameter family; a future measurement that pins the high-virtuality content of the triple-Pomeron vertex would sharpen or overturn the whole extrapolation.

The simultaneous fit to the total cross section and to $F_2$ at a single low scale is a consistency check rather than a prediction; the sharper test is whether the same parameters survive at a higher $Q^2$ where the semihard component dominates.

Since nucleus collisions reuse the same machinery with no new parameters, the predicted violation of superposition gives a cleaner discriminator than the $pp$ cross section, which was already well measured before this work.

The statement that inclusive jets stay factorizable while the eikonal does not suggests a useful diagnostic: observables built from AGK-cancelling combinations should be insensitive to the new screening, and any measured sensitivity would point to a breakdown of the soft-vertex assumption.

This review was written by a machine. No person has reviewed this paper on Pith yet. A scientist listed on Pith can review it for a fee, and their review is published here whatever it concludes.

Request a human review