{"id":"f06a430e-1731-4320-ae8d-8c9f6891cf5c","arxiv_id":"2608.04648","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Simulations show that low-flow sediment ridges enhance near-wall turbulence through a form-induced spectral peak, while stronger flows destroy the ridges and suppress turbulence.","lead":"This study uses computer simulations of grains in flowing water to show how the shape of the sand bed changes turbulence near the bed. The results could improve predictions of sediment transport in rivers and wind.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The streamwise form-induced second peak (λx+ ≈ 3220) is at or beyond the fundamental mode of the domain (Lx+ ≈ 3208), so the central spectral evidence is not a resolved peak and may be a finite-box artifact.","rationale":"The reader identified the fixed-uτ scaling as the weakest assumption; my concern sharpens it into a concrete inconsistency. The reported λx+ for the Sh1 form-induced second peak exceeds the streamwise box length computed from the paper's own parameters, so either the scaling is not actually by Sh0's uτ (violating the cross-case comparison) or the spectral maximum is pinned at the lowest wavenumber of the box and is not a real peak. This is load-bearing because the 'secondary peak exceeding the near-wall peak' is the paper's core mechanistic evidence; if that peak is a finite-domain artifact, the central claim of form-induced TKE enhancement is no longer supported by the spectral data. The spanwise peak remains resolved, but the streamwise claim and the 'non-monotonic modulation' story need to be re-established with a longer domain or with a demonstration that the fundamental mode is not the source. I therefore keep the conditional recommendation, but the condition is now specifically: show that the streamwise secondary peak is not a fundamental-mode artifact (e.g., by recomputing with per-case uτ and by checking the discrete wavenumber location). This goes beyond the reader's general uτ concern, so agreement is partial.","tokens_in":9162,"tokens_out":7865,"duration_ms":90093,"concrete_test":"Take the Sh1 velocity field used for Fig. 6(a3). Recompute the streamwise premultiplied spectrum using the per-case friction velocity (from the imposed pressure gradient and the measured H_b) and list the discrete wavenumbers; if the maximum occurs at the fundamental kx = 2π/Lx, the 'peak' is a boundary artifact. Independently verify the reported inconsistency: compute Lx+ = 13.63 × 8.77 × 26.84 ≈ 3208 and compare to the claimed λx+ = 3220. A definitive test is to run a second Sh1 case in a domain of Lx ≈ 27Hf (or at least twice as long) and check whether the streamwise secondary peak persists at the same λx+ or shifts with the box length; if it disappears or moves, the claimed mechanism is not scale-invariant.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The paper's central claim rests on the form-induced second peak in the streamwise premultiplied spectrum (Fig. 6a3), reported at λx+ ≈ 3220. However, from Table 1 and the stated Dp+ = 26.84, the streamwise box length is Lx+ = (Lx/Hf)(Hf/Dp)Dp+ ≈ 13.63 × 8.77 × 26.84 ≈ 3208. The reported peak is therefore at the largest resolved scale of the periodic domain—or, if exactly 3220, beyond it. This is not a converged spectral peak but a low-wavenumber pile-up, and its amplitude is controlled by the domain length rather than by the bedform physics. If the data were scaled with uτ of Sh0 as stated in §2, λx+ cannot exceed Lx+, revealing an internal inconsistency; if a per-case uτ was used, the comparison to Sh0 in common wall units is invalid, which is exactly the reader's concern. Either way, the quantitative claim that the streamwise secondary peak 'exceeds the conventional near-wall turbulent peak' is unsupported by resolved spectral data. The spanwise peak (λz+ ≈ 377) is well resolved, but the streamwise half of the headline claim and the associated TKE-enhancement mechanism depend on this unresolved mode.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports particle-resolved direct numerical simulations of open-channel flow over an erodible sediment bed at three Shields numbers (theta/theta_cr ≈ 2.0, 5.4, 8.4), together with a fixed-rough-bed baseline (Sh0) and a smooth-wall baseline (PF). Varying only the gravity magnitude to change the Shields number, the authors observe a morphological transition from streamwise sediment ridges (Sh1), through erosion trenches (Sh2), to transverse ripples (Sh3). The central claim is that this transition modulates near-wall turbulence non-monotonically: at low theta, form-induced streaks generated by the longitudinal ridges produce a secondary peak in the premultiplied energy spectra that exceeds the conventional near-wall peak and enhances turbulent kinetic energy; at higher theta, saltation disrupts these structures, the secondary peak disappears or weakens, and turbulence is suppressed. Proper orthogonal decomposition of the bed-height field is used to document a redistribution of modal energy from a single dominant longitudinal mode to higher-order modes, with longitudinal remnants persisting at high theta.","tokens_in":9406,"tokens_out":12974,"duration_ms":136851,"significance":"If the result holds, it would help reconcile contradictory reports in the literature: erodible beds can either enhance (Vowinckel et al. 2014) or suppress (Ji et al. 2013) near-wall turbulence, and the present study proposes the Shields number as a controlling parameter that changes both bedform orientation and saltation intensity. The study's strengths include the use of fully resolved particles with an immersed boundary method, a streamwise domain (13.63 H_f ≈ 120 D_p) that exceeds the threshold previously associated with transverse bedform onset, and the comparison of three Shields numbers while keeping fluid and particle properties fixed. The simulated bedload flux agrees with the Meyer-Peter-Müller and Wong-Parker correlations, which is a useful validation. However, the central spectral evidence is compromised by a finite-box inconsistency in the reported streamwise peak wavelength (see major comments), and the causal attribution to the bedform transition is not uniquely identified because the Shields number also changes the saltation intensity directly.","major_comments":[{"comment":"The reported streamwise form-induced second peak at lambda_x+ ≈ 3220 exceeds the fundamental streamwise period L_x+ ≈ 3208 of the periodic domain (using L_x/H_f = 13.63, H_f/D_p = 8.77, and D_p+ = 26.84 as stated in §2 and Table 1). A discrete Fourier mode in a periodic box cannot have a wavelength longer than L_x, so this value is internally inconsistent with the statement that all quantities are non-dimensionalized by u_tau of Sh0. The discrepancy likely arises from applying a per-case u_tau for Sh1–Sh3, but no Re_tau or u_tau is reported for the erodible cases. If a per-case u_tau was used, the comparison with Sh0 in common wall units is invalid; if the Sh0 u_tau was used, the reported lambda_x+ is impossible. The authors must report u_tau (or Re_tau) for Sh1–Sh3, recompute the spectra with a consistent normalization, and either demonstrate that the streamwise peak is a resolved feature independent of domain length or remove or reframe the claim that the streamwise secondary peak exceeds the conventional near-wall turbulent peak. The spanwise peak at lambda_z+ ≈ 377 is well resolved relative to L_z+ ≈ 802, but the streamwise half of the headline claim depends on the unresolved or mis-normalized mode.","section":"§3, Fig. 6(a3), Tables 1–2"},{"comment":"Each of the three erodible cases is a single realization with no error bars or uncertainty quantification. The non-monotonic modulation claim rests on differences in peak values (e.g., the streamwise second peak of 1.85 versus the first peak of 1.19 in Sh1, and the spanwise second peak decaying from 2.02 to 1.38 to 0.75 across Sh1–Sh3) and on differences in the fluctuation-profile maxima. The reported statistically steady durations are finite (T_obs^s/T_b = 258.68 for all cases). The authors should provide at least a split-sample convergence check for the spectra and fluctuation profiles, or otherwise quantify the sampling uncertainty, to show that the observed differences are not within the noise of time averaging.","section":"§3, Figs. 4 and 6; Tables 1–2"},{"comment":"The title and conclusion state that the turbulence modulation is driven by the bedform transition, but the simulations vary only the Shields number, which simultaneously changes saltation intensity and bed morphology. The observed correlation between bedform type and turbulence statistics is consistent with the proposed mechanism, but it does not separate the effect of bedform geometry from the direct effect of moving particles on the flow. To support the causal claim, the authors would need either a control with a fixed bed whose shape is prescribed (ridge versus ripple) at matched transport conditions, or a statistical decomposition that isolates the form-induced contribution while holding saltation statistics fixed. The current use of the time-averaged form-induced velocity e_bar_u is suggestive but is not a causal decomposition.","section":"§1, §3 (Fig. 5), §4"},{"comment":"The secondary spectral peak is labeled form-induced because of its location near the bed and its visual correspondence with the topography-conditioned streaks in Fig. 5, but the spectra in Fig. 6 are computed from the total velocity fluctuations, not from the form-induced velocity e_bar_u = u_f - <u_f>_xz. The authors do not show a spectrum of e_bar_u, nor a decomposition of the total spectrum into form-induced and turbulent parts. Without such a decomposition, the attribution of the peak to form-induced streaking remains a plausible interpretation rather than a demonstrated result. Presenting spectra of e_bar_u, or a scale-by-scale correlation with the bed-height fluctuation eta', would make the claim testable.","section":"§3, Figs. 5 and 6"}],"minor_comments":[{"comment":"The notation for temporal averaging is defined but not used subsequently; please either use it consistently or remove it.","section":"§2, Eq. (2.1)"},{"comment":"The definition of the premultiplied spectra (k_x+ E+ and k_z+ E+) and the spectral estimation details (windowing, binning, number of snapshots) are not provided; please add them to the methodology.","section":"§3, Fig. 6"},{"comment":"The POD implementation (snapshot method, normalization, number of modes retained, convergence) is not described, so the modal percentages are not reproducible.","section":"§3, Fig. 3"},{"comment":"The claim that the longitudinal bedforms at Sh1 are stable rests on a single realization; a longer observation window or an additional realization would make the claim more robust.","section":"§3, Fig. 2"},{"comment":"The definition of the form-induced velocity e_bar_u is ambiguous because the text calls it both a time-averaged field and the spatial fluctuation of u_f; please define it precisely with an equation.","section":"§3, Fig. 5"},{"comment":"The artificial blank region for Y+ < 0 in the PF panel is visually confusing; please explain it in the caption or remove it.","section":"§3, Fig. 6(a1,b1)"},{"comment":"The citation 'Y alin1977' appears to contain a typo; please check the citation and the corresponding reference list entry.","section":"§1"},{"comment":"The phrase 'for the first time' and the strong causal language ('directly link') are not fully supported by three simulations; consider tempering these claims.","section":"§1 and §4"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is within the journal's scope and the simulations appear carefully conducted, but the reported streamwise secondary spectral peak at lambda_x+ ≈ 3220 is inconsistent with the domain length L_x+ ≈ 3208 computed from the paper's own parameters. This is a load-bearing issue for the central claim, so I recommend major revision rather than rejection, because the spanwise spectral evidence and the flow visualization still qualitatively support a form-induced effect, and the authors may be able to correct the normalization and reframe the claim. Please ask the authors to report per-case friction Reynolds numbers and recompute the spectra with a consistent wall-unit normalization, and to address the statistical uncertainty and the confounding of bedform and saltation effects. The citation pattern and novelty disclosure appear appropriate."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: the systematic Shields-number sweep in a single simulation framework is genuinely new, and the bedform transition from streamwise ridges through erosion trenches to transverse ripples is a useful result. The POD 'top-left to bottom-right' redistribution is a nice way to visualize the transition, and the particle-flux agreement with Meyer-Peter-Mueller and Wong-Parker adds credibility. The fluctuation profiles show a clear non-monotonic response relative to the fixed-bed baseline, independent of the spectral interpretation. That part is worth reading.\n\nThe problem is the spectral centerpiece. The streamwise 'form-induced second peak' is reported at λx+ ≈ 3220. With the paper's own Sh0 scaling, the streamwise box length is Lx+ ≈ 13.63 × 8.77 × 26.84 ≈ 3208. So the claimed peak is at, or slightly beyond, the largest scale the periodic domain can represent. That is not a resolved spectral peak; it is the fundamental-mode pile-up. The authors even say the structure 'nearly spans the entire domain in Sh3', which points to the same issue. Either the scaling used a different friction velocity for Sh1, which breaks the comparison to Sh0, or the number is simply inconsistent. Either way, the abstract's claim that this peak 'exceeds the conventional near-wall turbulent peak' is not backed by well-resolved spectral data.\n\nThis is load-bearing: the TKE-enhancement mechanism is tied to that peak, and the spanwise peak alone, while resolved, is not enough to carry the full argument. The paper needs a longer box, a convergence check, or a reworded interpretation that does not claim a resolved peak.\n\nSecondary soft spots: using u_tau of Sh0 for all erodible cases requires justification, but none is given and no Re_tau is reported for Sh1–Sh3. Each case is a single realization with no uncertainty quantification. The POD-to-turbulence link is correlational. These are secondary but should be addressed in revision.\n\nOverall: worth sending to a serious referee, but the spectral claim needs major rework before publication. The bedform-transition results and the fluctuation data are solid enough to keep the paper alive.","headline":"Solid PR-DNS study of bedform transitions, but the headline streamwise spectral peak sits at the periodic-domain fundamental mode, so the central mechanism claim is not yet supported.","tokens_in":9934,"tokens_out":4807,"would_cite":false,"duration_ms":52832,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper claims that over an erodible bed, raising the Shields number turns streamwise sediment ridges into transverse ripples and, in doing so, first enhances near-wall turbulence through form-induced streaks and a secondary spectral…","keywords":["erodible bed","Shields number","bedform transition","turbulence modulation","premultiplied energy spectra","form-induced streaks","particle-resolved DNS","proper orthogonal decomposition"],"falsifier":"Compute the friction velocity $u_\\tau$ independently for each erodible case (Sh1-Sh3) rather than reusing Sh0's value; if the friction Reynolds number changes measurably between the cases, the reported $Y^+$ locations of the form-induced second peak shift, and after the correction the secondary peak may no longer exceed the conventional near-wall peak. Equivalently, a spectral measurement at $Y^+ \\approx 5$ with $\\lambda_x^+ \\approx 3220$ in a different apparatus at the same $\\theta/\\theta_{cr}$ would confirm or refute the universal role of the form-induced peak.","tokens_in":8932,"feed_emoji":"🌀","tokens_out":9358,"duration_ms":91281,"temperature":0.7,"pith_summary":"The paper uses particle-resolved direct numerical simulations of an open channel with an erodible sediment bed to show that near-wall turbulence is modulated non-monotonically as the Shields number increases. At low Shields numbers, streamwise sediment ridges act as form-induced roughness that creates a distinct secondary peak in the premultiplied energy spectra near the wall, a peak that exceeds the conventional turbulent near-wall peak and raises turbulent kinetic energy relative to a fixed rough bed. As the Shields number increases, intensified saltation disrupts these ridge-induced streaks, making the secondary peak vanish in the streamwise direction and weaken in the spanwise direction, so turbulence is suppressed. The point of the claim is that conflicting experimental results on whether mobile beds enhance or suppress turbulence can be reconciled by where each flow sits on the Shields-number axis.","feed_headline":"Erodible ridges add a spectral peak; saltation erases it","feed_subtitle":"A second near-wall energy peak appears with sediment ridges and fades as saltation takes over.","key_machinery":"The controlling parameter is the Shields number $\\theta = \\tau_w/(\\rho_f R |\\boldsymbol{g}| D_p)$, with $R = \\rho_p/\\rho_f - 1$, which sets the balance between fluid shear and immersed weight and is varied here by changing gravity alone. The diagnostic that carries the argument is the one-dimensional premultiplied energy spectrum, $k_x^+ E_{uu}^+$ and $k_z^+ E_{ww}^+$, whose area represents turbulent energy at each scale and whose wall-normal evolution reveals a secondary peak at $Y^+ \\approx 1$ to $5$ when streamwise ridges are present. The form-induced velocity field $\\bar{u} = u_f - \\langle u_f\\rangle_{xz}$ isolates the spatial, time-averaged velocity footprint of the bed topography and identifies topography-conditioned streaks. Proper orthogonal decomposition of the bed-surface fluctuation $\\eta'$ supplies the modal picture that links bedform morphology to the turbulence statistics.","core_discovery":"The central discovery is that the transition of an erodible bed from streamwise ridges to transverse ripples, driven solely by increasing the Shields number via gravity, changes the near-wall energy budget through a competition between static bed particles and moving saltating particles. Low-Shields ridges generate topography-conditioned streaks and a form-induced second peak in the premultiplied spectra (streamwise peak $k_x^+ E_{uu}^+ = 1.85$ at $Y^+ = 1$, $\\lambda_x^+ = 3220$; spanwise peak $k_z^+ E_{ww}^+ = 2.02$ at $Y^+ = 3$, $\\lambda_z^+ = 377$) that exceeds the conventional near-wall turbulent peak and enhances TKE relative to the fixed rough bed. With rising Shields number, saltation erodes and reshapes the bed, the streamwise secondary peak disappears completely, the spanwise peak weakens, and all three velocity fluctuation components are suppressed relative to Sh0. Proper orthogonal decomposition of the bed-height field shows a top-left to bottom-right redistribution from a single dominant longitudinal mode to higher-order modes, with quasi-longitudinal remnants surviving as the morphological signature of the transition.","pith_inferences":["Editorial: if the wall-unit scaling from Sh0 is not valid for Sh1-Sh3, because moving particles change the effective roughness and hence $u_\\tau$, the reported $Y^+$ locations of the spectral peaks would shift; reporting $Re_\\tau$ for each erodible case would directly test this.","Editorial: the same non-monotonic modulation might be collapsed across particle sizes by comparing the ratio of saltating particle flux to the form-induced velocity amplitude in the wall layer rather than the Shields number alone.","Editorial: because the streamwise secondary peak vanishes while a weakened spanwise peak persists at higher Shields numbers, spanwise spectra should be more sensitive diagnostics of residual bedform influence in aeolian and fluvial field measurements.","Editorial: the POD residual-longitudinal-mode signature suggests a snapshot-based classifier: even when transverse ripples dominate, the presence of quasi-longitudinal high-order modes marks the bed as being in transition rather than fully washed out."],"forward_implications":["If the mechanism is right, the same bed can either enhance or suppress near-wall turbulence depending on transport stage, so conflicting literature reports should correlate with where each experiment sits on the Shields-number axis.","Longitudinal sediment ridges can remain stable at lower Shields numbers even in domains long enough that previous studies predicted a transition to transverse bedforms.","The bedform sequence with increasing Shields number, from depositional ridges through erosional trenches to transverse ripples, provides a morphological ordering that can be read directly from bed-height snapshots.","The form-induced second peak acts as a quantitative fingerprint: when it appears at $Y^+ \\approx 5$ and exceeds the conventional peak, the bedform configuration is enhancing turbulence, and when it vanishes or drops below, saltation has taken over.","The POD top-left to bottom-right transfer of modal weight implies that the destruction of coherent longitudinal bedforms by saltation is what drives the turbulence suppression, not the mere presence of mobile particles."],"supporting_citations":[{"why":"Established the long-domain threshold beyond which longitudinal bedforms give way to transverse bedforms; the paper's Sh1 persistence at lower theta is a direct point of comparison.","marker":"Kidanemariam & Uhlmann (2017)"},{"why":"Simulated streamwise sediment ridges and showed they disrupt the near-wall self-sustaining cycle; supplies the baseline for Sh1's suppressed streamwise peak.","marker":"Scherer et al. (2022)"},{"why":"Particle-resolved DNS of an erodible bed that reported suppression of all fluctuation peaks at similar theta over theta_cr; the discrepancy with Sh1 shows bedforms can reverse that suppression.","marker":"Ji et al. (2013)"},{"why":"Found ridge-like and dune-like bedforms enhance Reynolds stress components; used to interpret why the present truly erodible bed behaves differently.","marker":"Vowinckel et al. (2014)"},{"why":"Supplies the empirical critical Shields number theta_cr used to normalize all four cases.","marker":"Guo (2020)"},{"why":"Classic bed-load transport formula used to validate the particle volume-flux curve in figure 1(b).","marker":"Meyer-Peter & Müller (1948)"},{"why":"Reanalyzed bed-load database; its correction of the Meyer-Peter-Muller formula is the second validation curve for the flux data.","marker":"Wong & Parker (2006)"},{"why":"Provides the second-order immersed boundary method that carries the fluid-particle coupling in the simulations.","marker":"Breugem (2012)"},{"why":"Collision model with adaptive collision time used for particle-particle and particle-wall contacts.","marker":"Biegert et al. (2017)"}],"fun_headline_variants":["Ridges boost near-wall energy, saltation kills it","Shields number flips bedforms and turbulence","From ridges to ripples: turbulence peaks fade","Saltation wipes out the ridge-driven energy peak"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The comparison assumes that changing only gravity, with the same pressure gradient, leaves the friction velocity and the wall-unit coordinates effectively unchanged across all erodible cases, even though moving sediment should alter the bed's effective roughness.","fun_headline_variants_meta":{"raw":{"variants":["Ridges boost near-wall energy, saltation kills it","Shields number flips bedforms and turbulence","From ridges to ripples: turbulence peaks fade","Saltation wipes out the ridge-driven energy peak"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000778,"raw_usage":{"total_tokens":3439,"prompt_tokens":947,"completion_tokens":2492,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":563,"completion_tokens_details":{"reasoning_tokens":2429}},"tokens_in":563,"tokens_out":2492,"duration_ms":19458,"temperature":1.0,"reasoning_tokens":2429,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T19:51:18.733789+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the friction velocity $u_\\tau$ independently for each erodible case (Sh1-Sh3) rather than reusing Sh0's value; if the friction Reynolds number changes measurably between the cases, the reported $Y^+$ locations of the form-induced second peak shift, and after the correction the secondary peak may no longer exceed the conventional near-wall peak. Equivalently, a spectral measurement at $Y^+ \\approx 5$ with $\\lambda_x^+ \\approx 3220$ in a different apparatus at the same $\\theta/\\theta_{cr}$ would confirm or refute the universal role of the form-induced peak.","supporting_citations":[{"cited_title":"& Uhlmann, Markus 2017 Formation of sediment patterns in channel flow: minimal unstable systems and their temporal evolution","cited_arxiv_id":null,"evidence_quote":"Established the long-domain threshold beyond which longitudinal bedforms give way to transverse bedforms; the paper's Sh1 persistence at lower theta is a direct point of comparison."},{"cited_title":"& Krayer, Michael 2022 On the role of turbulent large-scale streaks in generating sediment ridgesa","cited_arxiv_id":null,"evidence_quote":"Simulated streamwise sediment ridges and showed they disrupt the near-wall self-sustaining cycle; supplies the baseline for Sh1's suppressed streamwise peak."},{"cited_title":"Advances in Water Resources 72 , 32--44","cited_arxiv_id":null,"evidence_quote":"Found ridge-like and dune-like bedforms enhance Reynolds stress components; used to interpret why the present truly erodible bed behaves differently."},{"cited_title":"Journal of Hydraulic Engineering 146 (6), 4020038","cited_arxiv_id":null,"evidence_quote":"Supplies the empirical critical Shields number theta_cr used to normalize all four cases."},{"cited_title":"& Müller, R","cited_arxiv_id":null,"evidence_quote":"Classic bed-load transport formula used to validate the particle volume-flux curve in figure 1(b)."},{"cited_title":"Journal of Computational Physics 231 (13), 4469--4498","cited_arxiv_id":null,"evidence_quote":"Provides the second-order immersed boundary method that carries the fluid-particle coupling in the simulations."},{"cited_title":"Journal of Computational Physics 340 , 105--127","cited_arxiv_id":null,"evidence_quote":"Collision model with adaptive collision time used for particle-particle and particle-wall contacts."}],"review_version":1}