Improved low Q2 model for neutrino and electron nucleon cross sections in few GeV region

This is the NuInt04 proceedings version of the Bodek–Yang effective LO model, and the honest summary is: the charged-lepton sector is solid, the neutrino sector is not yet demonstrated, and the paper's own numbers don't quite agree with each other. Worth reading if you work on few-GeV neutrino cross sections; not worth treating as a finished result. What's actually new relative to hep-ex/0308007 is modest but real: the low-Q2 K factors are split into separate u and d valence and sea pieces, and the proton/deuteron photoproduction data are moved from a prediction into the fit. The global fit to SLAC/BCDMS/NMC/H1 F2 plus photoproduction gives chi2/DOF = 1874/1574, which is fine, and the resonance comparison in Figure 4 is a genuine out-of-sample check — the model tracks the average resonance F2 down to Q2 = 0.07 via duality. That's the part I'd trust. Two soft spots, one bigger than the other. The bigger one is reproducibility: the text on p.2 quotes A=0.538, Cv1d=0.202, Cv2d=0.255, fPDF=1.015, while the Figure 1 and 4 captions quote A=0.541, Cv1d=0.227, Cv2d=0.286, fPDF=1.014, plus aLambda=2.052 and a set of dataset normalizations that never appear in the text. The d-valence numbers differ by roughly 10%. For a paper whose deliverable is a parameter set, that matters — the quoted chi2 isn't tied to a unique fit, and nothing in the text acknowledges the discrepancy. No parameter uncertainties or covariance are given either, so you can't judge how well the d-valence split is actually determined. The smaller one is the neutrino claim. The authors are upfront that the CCFR/CDHSW comparison assumes the axial-vector contribution equals the fitted vector one, and they flag the real treatment as future work in Eq. 6. So the body is honest; it's the abstract's 'describes ... neutrino data' that overreaches. Also note photoproduction is no longer an independent test now that it's in the fit. None of this is load-bearing for the charged-lepton model. It's a proceedings, so some of this is expected, but the parameter mismatch should be fixed before anyone uses these numbers. I'd cite it for the low-Q2 F2 parameterization and wait for the full paper on the axial-vector side.

The paper's central deliverable is the fitted parameter set that supposedly describes all the data. The text (p.2) quotes A=0.538, B=0.305, Cv1d=0.202, Cv1u=0.291, Cv2d=0.255, Cv2u=0.189, Cs1d=0.621, Cs1u=0.363, fPDF=1.015, chi2/DOF=1874/1574. The Figure 1 and Figure 4 captions quote A=0.541, B=0.306, Cv1d=0.227, Cv1u=0.291, Cv2d=0.286, Cv2u=0.189, Cs1d=0.619, Cs1u=0.363, fPDF=1.014, plus aLambda=2.052 and normalization factors nSLACd=0.990, nBCDMSp=0.967, nBCDMSd=0.987, nNMCp=1.003, nNMCd=0.994, nH1p=0.978, nH1d=1.000. The d-valence parameters differ by roughly 10%, and the text's list omits aLambda and the normalization factors entirely. So the quoted chi2/DOF cannot be attributed to a unique, reproducible parameter set, and the figures may have been produced with a different fit than the text describes. This is not acknowledged anywhere. By contrast, the reader's neutrino concern is explicitly acknowledged: the authors write 'As a preliminary step, we compare the CCFR and CDHSW energy neutrino data with our model, assuming that the vector contribution is same as the axial vector contribution,' and they flag the axial-vector treatment as future work in Eq. 6. That is an overclaim in the abstract (s3) relative to the body, but it is a stated limitation, not an unresolved internal inconsistency. The parameter mismatch is the unresolved one and it bears directly on whether the central claim is reproducible.

The reader's verdict (CONDITIONAL) is right, and the reader correctly notices both the neutrino vector=axial assumption and the text-vs-caption parameter inconsistency. I disagree only on which is the load-bearing unresolved concern. The neutrino assumption is acknowledged in the body ('As a preliminary step... assuming that the vector contribution is same as the axial vector contribution') and explicitly deferred to future work via the Adler-sum-rule K factors of Eq. 6. Per the review rules, an acknowledged limitation should not be presented as a new finding; the abstract's s3 claim to describe neutrino data is an overclaim relative to the body, but the body is honest about it. The parameter mismatch, by contrast, is nowhere acknowledged: the text's parameter list and the figure captions disagree on A, B, Cv1d, Cv2d, Cs1d, and fPDF, and the captions include aLambda and seven normalization factors that the text omits. For a paper whose product is a parameterization, that is the more load-bearing gap because it makes the quoted chi2/DOF and the figures non-reproducible from the paper alone. The charged-lepton fit itself (chi2/DOF=1874/1574, acceptable) and the out-of-fit resonance comparison in Figure 4 are legitimate; the photoproduction-in-fit point the reader raises is a validation gap, not an error. So the verdict stays CONDITIONAL, but the condition should be stated as: reconcile the text and caption parameter sets and state which one produced the figures and the quoted chi2.

If the model holds, it is a genuinely useful object: a small parameter set that interpolates continuously from the photoproduction limit (Q^2 = 0) through the resonance region to DIS, and that can be transported from precise charged-lepton data to neutrino cross sections. Two features deserve explicit credit. First, the photoproduction data are now included as fit constraints rather than merely predicted, which is the right way to anchor the Q^2 -> 0 behaviour of the K factors. Second, the comparison with JLab resonance data in Fig. 4 is an out-of-fit test of the duality-based averaging, and it is a falsifiable prediction of the model rather than a post-hoc description. The separate treatment of u and d valence and sea K factors is a modest but real improvement over the single-K-factor version, and the resulting parameterization is directly usable by oscillation experiments that must model few-GeV cross sections.

The paper asks whether one effective leading-order quark-parton model can cover the whole kinematic span that few-GeV neutrino experiments actually traverse: deep inelastic scattering at high $Q^2$, the resonance region, and the photoproduction limit at $Q^2=0$. The authors' tool is a rescaled momentum fraction $xi_w$ that absorbs target-mass and higher-twist effects, combined with multiplicative $K$ factors that suppress the quark distributions at low $Q^2$ differently for up valence, down valence and sea quarks. They fit this machinery to all inelastic charged-lepton $F_2$ data (SLAC, BCDMS, NMC, H1) together with photoproduction cross sections on proton and deuteron, and report $chi^2/mathrm{DOF} = 1874/1574$. The payoff, if the fit holds, is that a single parameterization supplies the inelastic part of the neutrino cross section in the few-GeV region, where direct neutrino data are sparse and where oscillation experiments need reliable predictions. The authors also check the model against Jefferson Lab resonance data, which were not in the fit, and against CCFR and CDHSW neutrino data on iron.

What carries the argument. Two pieces carry the argument. The first is the scaling variable $xi_w = 2x(Q^2 + M_f^2 + B) / left(Q^2left[1 + sqrt{1 + (2Mx)^2/Q^2}right] + 2Axright)$, which descends from energy-momentum conservation with a target-mass denominator and two free parameters: $A$ standing in for dynamical higher-twist terms as an enhanced target-mass correction, and $B$ standing in for initial-state transverse momentum and the effective final-state mass shift from multi-gluon emission, which is what allows the model to reach the $Q^2 = 0$ photoproduction limit. The second is the set of multiplicative $K$ factors, $K_{sea}(Q^2) = Q^2/(Q^2 + C_s)$ and $K_{valence}(Q^2) = [1 - G_D^2(Q^2)] cdot (Q^2 + C_{v2})/(Q^2 + C_{v1})$, with $G_D$ the proton elastic form factor; the factor $[1 - G_D^2]$ behaves as $Q^2/(Q^2 + 0.178)$ at low $Q^2$ and forces the structure function toward the photoproduction cross section through $sigma(gamma p) = 4pi^2alpha_{EM} F_2/Q^2$. Evolution of the GRV98 distributions is frozen below $Q^2 = 0.8$ GeV$^2$, and the low-$Q^2$ structure function is obtained by evaluating the frozen distributions at $xi_w$ and applying the $K$ factors.

Core claim. The claim is that an effective leading-order model, formed by taking the GRV98 LO parton distributions, replacing $x$ with the variable $xi_w$, and multiplying the quark distributions by $Q^2$-dependent $K$ factors applied separately to $u$ valence, $d$ valence and sea quarks, reproduces the measured inelastic structure function over the full range from high $Q^2$ down to $Q^2 = 0$. The fit uses charged-lepton $F_2$ data from SLAC, BCDMS, NMC and H1 plus photoproduction data on proton and deuteron, and returns $A = 0.538$, $B = 0.305$, $C_{v1d} = 0.202$, $C_{v1u} = 0.291$, $C_{v2d} = 0.255$, $C_{v2u} = 0.189$, $C_{s1d} = 0.621$, $C_{s1u} = 0.363$ and an overall PDF rescaling $f_{PDF} = 1.015$, with $chi^2/mathrm{DOF} = 1874/1574$. The model, without further tuning, lands on the average level of the Jefferson Lab resonance data and on the CCFR and CDHSW neutrino cross sections on iron. The authors present this as an improvement on their earlier version, in which a single $K$ factor covered all quark species; splitting it by flavour and by valence versus sea is what lets photoproduction data enter the fit as a constraint rather than as a post-hoc prediction.

The premise it rests on. The neutrino part of the claim rests on comparing the model to neutrino data while assuming the axial-vector structure function equals the vector one, an equality the authors themselves say does not hold at low $Q^2$ and which they are still working to constrain with Adler-sum-rule-motivated corrections.

A single parameterization covers deep inelastic scattering, the average of the resonance region, and photoproduction, so few-GeV neutrino cross sections can be computed without stitching together separate models at each kinematic boundary.

Because the corrections are applied per quark species, the ratio of up to down structure functions at low $Q^2$ is no longer fixed by the high-$Q^2$ distributions, which changes predicted cross sections on neutron-rich and proton-rich nuclear targets.

The agreement with Jefferson Lab resonance data, which were not fitted, supports the duality picture in which an average over resonances already tracks the deep inelastic curve for the same $xi_w$.

Neutrino oscillation analyses in the few-GeV region gain a systematically constrained inelastic background estimate, which directly affects the extracted oscillation parameters through the predicted event rates.

The same $K$-factor structure, once extended to the axial-vector channel, would give a unified charged-current description covering quasi-elastic, resonance and inelastic contributions.

The neutrino comparison in the paper assumes the axial-vector structure function equals the vector one; since the two are known to differ at low $Q^2$, the agreement shown there is partly built in, and the model's real neutrino test will come only after the Adler-sum-rule-motivated axial $K$ factors of Eq. 6 are fitted.

Because the low-$Q^2$ suppression is fitted separately for $u$ and $d$ valence quarks, the model predicts a definite $Q^2$ dependence for the ratio of proton to deuteron structure functions in the resonance region; a precise measurement of that ratio at fixed $xi_w$ would test the flavour split independently of the overall normalization.

The frozen-evolution prescription below $Q^2 = 0.8$ GeV$^2$ means the model has no genuine $Q^2$ dependence there beyond the $K$ factors, so any observed structure in the low-$Q^2$ slope of $F_2$ that is not captured by the form-factor-inspired factor would show up as a discrepancy.

If the model holds, the same machinery could be reused for electron-nucleus scattering in the few-GeV range, where the resonance-to-continuum transition affects the interpretation of inclusive quasielastic measurements.

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