{"id":"a967137b-8200-4df1-b7fd-65f56120154d","arxiv_id":"2411.13406","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"A DNS-calibrated vortex-wake model added to an agent-based fish school predicts that wakes impose oblique, diamond-like order on high-alignment schools.","lead":"This paper adds a computer-modeled fish wake, tuned to high-resolution flow simulations, into an agent-based model of schooling fish. It predicts that wake vortices push schools into more ordered, diagonal formations, especially when fish strongly align with their neighbors.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The organizing effect of wakes rests on unvalidated linear superposition: the diamond pattern requires wake-edge attraction that may vanish when wakes are modified by neighbors. A two-fish DNS check can settle it.","rationale":"Read in good faith: the model is a deliberate phenomenological framework, the wake parameters are fitted to single-fish DNS, and the school-level patterns are emergent rather than fitted, so the comparison is not circular. The paper also provides useful ensemble statistics. However, the central claim is not merely 'the model produces organized patterns' but that vortex wakes improve organization; that requires the wake representation to be adequate in the multi-fish regime. The explicit superposition/no-modification assumption is the weakest link because the mechanism proposed for the diamond pattern is built on it. The proposed DNS check is feasible given the authors' existing solver and would directly test the wake-edge attraction. The quasi-steady force coefficient issue compounds the concern but is secondary. Thus the verdict should remain conditional, with the added condition that the wake superposition be validated in a multi-fish configuration or the claim be re-framed as a prediction of the idealized model. This agrees with the reader's identified weakest assumption.","tokens_in":17235,"tokens_out":6851,"duration_ms":82744,"concrete_test":"Use the same ViCar3D solver to simulate a two-fish configuration with the leading fish swimming steadily and the trailing fish held at the model's predicted wake-edge offset (the diamond position in Fig. 12, e.g., approximately 0.5 LB lateral and 1 LB downstream). Compute the time-averaged surge force, sway force, and yaw moment on the trailing fish, and compare them to the forces predicted by the model from Eq. 4 using the linear superposition of the leading fish's single-fish wake. Also compare the wake velocity field around the trailing fish to the linear superposition of isolated-fish wakes. If the predicted favorable forward-drafting region at the wake edge is absent, reversed, or shifted by more than a body-width, the wake-induced diamond organization is an artifact of the superposition assumption. Repeating for 3 fish tests the 'enhanced edge' feedback.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"Section II.H explicitly assumes that the wake of each fish is not modified by trailing or adjacent fish, and Section II.E represents the school flow as a linear sum of single-fish potential and Rankine-vortex wakes. The central result in Section III.B — the emergence of a streamwise diamond pattern under high alignment — is explained mechanistically by superposed wake edges: 'the wake region behind the trailing fish would consist of one superposition of their wake's right edges and two left edges... the enhanced right edge would have a stronger induced velocity to attract the following trailing fish.' Thus the improvement in PCA-based organization (11.95% vs 1.54% first-PC variance) is not a robust physical prediction but a consequence of assuming that wake vortices pass through and around neighbors unmodified. In real multi-fish flows, the presence of a trailing fish alters the pressure and vorticity field, and vortices can be stretched, merged, or deflected; the favorable drifting region that the mechanism depends on may shift or disappear. The force coefficients in Eq. 4 are also quasi-steady (static-body DNS at angle of attack), so the response of a fish to an unsteady periodic wake is not captured. The central claim would be settled by validating the superposition against multi-fish DNS.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a two-dimensional agent-based model of fish schooling in which each fish is subject to social forces (attraction, alignment, wall avoidance) and to hydrodynamic forces from a potential-flow representation of the body plus a discrete Rankine-vortex wake whose parameters are fitted to the authors' three-dimensional direct numerical simulations of a single carangiform swimmer. The model is used to compare school topology in simulations with and without the wake model, for various numbers of fish and for high-alignment (alpha_T=0.1) versus high-attraction (alpha_T=0.9) social regimes. The central claim is that adding the vortex wake increases the spatial organization of the school, most strongly in the high-alignment regime, producing an oblique 'diamond' pattern; the first principal component of the school distribution explains 11.95% of the variance with wake versus 1.54% without (Section III.B, Fig. 15). The paper also reports sensitivities to wake angle (Section III.C) and vortex strength (Section III.D).","tokens_in":17574,"tokens_out":6573,"duration_ms":64299,"significance":"If the central claim holds, the work is a valuable contribution because it bridges behavior-focused agent-based models and hydrodynamic realism, and it generates concrete, falsifiable predictions: e.g., wider wakes and stronger vortices should enhance schooling organization in highly polarized groups (Figs. 17 and 18). The use of 300 ensemble runs per condition and DNS-calibrated hydrodynamic coefficients is a strength, and the model's explicit separation of social and hydrodynamic mechanisms allows the wake effect to be isolated cleanly. However, the absence of multi-fish validation of the wake-superposition assumption and the lack of uncertainty quantification on the PCA metrics currently limit the strength of the physical conclusions. The paper is well-suited to the journal and would benefit from targeted additional validation.","major_comments":[{"comment":"Section II.H states that 'the wake flow generated by a given fish is not modified by any trailing or adjacent fish,' and Section II.E (Eq. 8) represents the school flow as a linear superposition of single-fish potential and wake fields. The central mechanism for the diamond pattern in Section III.B — that wake edges from successive fish superpose to create stronger attractive edges — is a direct consequence of this assumption. Because the superposition has not been validated for multi-fish configurations, the physical claim that vortices improve school organization is not yet robust. I recommend adding a targeted validation (e.g., a two-fish DNS or experiment) or, at minimum, a sensitivity test that perturbs or attenuates the wake behind a trailing fish to show that the qualitative result does not depend on the exact superposition.","section":"Section II.H"},{"comment":"The hydrodynamic force and moment coefficients C1, C2, C3 are quasi-steady, obtained from DNS of a stationary fish at fixed angles of attack. In the school simulations, fish are subjected to a periodically unsteady wake from neighbors, and the quasi-steady approximation neglects unsteady effects such as added mass, wake history, and dynamic stall. These effects could alter the magnitude and location of the hydrodynamic forces that are claimed to attract trailing fish to the wake edge. Please justify the quasi-steady assumption quantitatively, for example by comparing with an unsteady estimate of the forces on a fish in a periodic wake, or by showing that the time-averaged forces dominate the dynamics.","section":"Section II.D, Eq. (4)"},{"comment":"The quantitative claim that wakes organize schools is based on the explained variance of the first principal component (11.95% vs. 1.54%). The PCA is computed from 300 simulations with five snapshots each, but no confidence intervals or statistical significance tests are reported. Since the ensemble size is finite, the difference could be sensitive to snapshot selection or to a few outlier configurations. Please include a bootstrap or permutation-based uncertainty estimate for the PCA variances, or otherwise demonstrate that the difference is statistically robust.","section":"Section III.B, Figs. 14 and 15"}],"minor_comments":[{"comment":"The formula for the flow angle phi appears mis-typeset; please clarify using an explicit atan2 expression for the two components of the relative velocity.","section":"Section II.D"},{"comment":"The preferred distance R0 is listed as '1/LB', which is dimensionally inconsistent; since lengths are nondimensionalized by LB, R0 should be a dimensionless value (or state R0 = 1 LB before nondimensionalization).","section":"Table II"},{"comment":"The segmentation of the vision field into six sectors is stated to be based on 'various tests' without details; please provide a reference, a description of the tests, or a brief sensitivity analysis.","section":"Section II.F"},{"comment":"The notation 'N = 1,500' in the captions refers to the number of PCA samples (300 simulations x 5 snapshots), which is easily confused with the number of fish; please relabel as 'Nsamples = 1,500'.","section":"Figs. 17 and 18 captions"},{"comment":"The description of the effect of wake angle on schooling organization is contradictory at first reading ('wider wakes greatly contribute' followed by 'wider wakes have a detrimental effect' for beta <= 11 degrees). Please revise to state the non-monotonic dependence clearly and define the threshold beta approx 14 degrees.","section":"Section III.C"},{"comment":"There is a typo 'Talbe I' in Section II.A, and in the Fig. 1 caption 'theta_ij is the angle between r_ij and U1' should refer to the surge direction rather than the scalar velocity U1.","section":"Section II.A and Fig. 1 caption"}],"recommendation":"major_revision","confidential_remarks":"The manuscript relies heavily on the authors' own prior DNS work (Refs. [13], [43]) and several self-citations to bulletins/APS abstracts; this is legitimate calibration, but the reliance on non-archival sources for the wake model details could be strengthened. The fit with Physics of Fluids is reasonable given the DNS parameterization, but the lack of multi-fish validation of the superposition assumption may be a concern for the readership; the proposed two-fish DNS check would substantially increase confidence in the central claim."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Punchline: this is a genuinely new modeling contribution that puts DNS-calibrated vortex wakes into an agent-based fish school model, and it shows wakes can sharpen school structure. But the headline diamond pattern depends on the assumption that wakes pass through neighbors unmodified, so take the quantitative PCA numbers as a model prediction, not a measured effect.\n\nWhat's new: prior schooling models either ignore hydrodynamics or use potential-flow dipoles. Here the wake is a discrete set of Rankine vortices arranged in an oblique pattern, with circulation and geometry fit to the authors' own 3D DNS of a carangiform swimmer. That's a real step up in physical content, and the force coefficients in the surge-sway-yaw equations also come from DNS. The ensemble methodology (300 runs per case, PCA on 1500 snapshots) is solid and appropriate for a nonlinear dynamical system.\n\nThe good: the paper is honest about the assumption that a fish's wake is not modified by trailing or adjacent fish (Sec. II.H). It also shows the with-wake vs without-wake comparison under identical social parameters, so the central comparison is not circular. The observation that wakes matter most when alignment dominates attraction is physically plausible and gives a testable handle.\n\nThe soft spots: the superposition and unmodified-wake assumption is the load-bearing simplification. The mechanism that creates the diamond pattern is explicitly the enhanced induced velocity from overlapping wake edges; in real multi-fish flows those vortices could merge, stretch, or deflect, and the favorable drift region might shift. A two-fish DNS would settle this. Also, the force coefficients are quasi-steady from static-body DNS at angle of attack, so unsteady forcing from an oscillating wake is not captured. There is no uncertainty quantification on the PCA variance fractions; five snapshots per run is small. The efficiency language in the conclusion goes beyond what the model computes – the model shows drift regions but never calculates muscle work or cost of transport. Finally, code and data are only 'available upon request,' which limits reproduction.\n\nWho it's for: researchers in collective behavior, bio-inspired underwater robotics, and fish locomotion modeling. It's a useful framework to build on. I'd send it to peer review, but I'd ask the referee to require either a validation of the superposition assumption or a frank about its consequences, plus UQ on the PCA metrics. If the wake-neighbor interaction turns out to kill the effect, the framework still has value as a tool, just with a weaker headline.","headline":"A DNS-anchored vortex-wake ABM for fish schools that shows wakes can order schools, but the headline pattern hinges on an unvalidated wake-superposition assumption.","tokens_in":18021,"tokens_out":2853,"would_cite":false,"duration_ms":30058,"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":"Simulations with a DNS-parameterized wake model show that vortex wakes organize fish schools into oblique diamond patterns, especially when social alignment dominates attraction.","keywords":["fish schools","agent-based model","hydrodynamic wakes","vortex wake model","direct numerical simulation","collective behavior","schooling topology","Rankine vortices"],"falsifier":"Simulate a small school (two or three fish) with the same carangiform kinematics using a fully resolved multi-body DNS and measure the induced velocity field behind the leading fish; if the trailing fish do not preferentially settle near the wake-edge drifting regions, or if the first-PC explained variance does not increase when the wake is present, the model's central organizing mechanism is falsified. More directly, compare the modeled oblique Rankine-vortex wake to the actual wake of a fish swimming in a school: if the wake is substantially modified by neighbors, the superposition assumption breaks.","tokens_in":17038,"feed_emoji":"🐟","tokens_out":5187,"duration_ms":53284,"temperature":0.7,"pith_summary":"This paper develops an agent-based model of fish schooling that adds a DNS-parameterized vortex wake to the usual social rules of attraction, alignment, and vision. The central claim is that hydrodynamic wakes are not a source of disorder but an organizing force: in simulations with strong alignment and weak attraction, turning on the wake changes the school from a diffuse, disorganized cluster into a streamwise-oblique diamond pattern, with the first principal component of the school's spatial distribution explaining 11.95% of the variance versus 1.54% without the wake. The authors argue that trailing fish are passively drawn to the narrow drifting regions near the edges of a leading fish's wake, and that this effect strengthens when alignment keeps fish consistently oriented. If correct, the result implies that school topology can encode hydrodynamic information and that purely social rules are insufficient to explain observed formations.","feed_headline":"Vortex wakes pull fish schools into ordered diamond patterns","feed_subtitle":"A DNS-parameterized model shows that wakes, not social rules alone, set the staggered formations of schooling fish.","key_machinery":"The central object is the phenomenological wake model: each half tail-beat sheds a line of discrete Rankine vortices of alternating sign, arranged in the oblique pattern observed in the authors' three-dimensional direct numerical simulations of a mackerel-like carangiform swimmer, with vortex strength and decay rates fitted to those simulations. This wake field is added to a potential-flow model of the fish body (four source-sink pairs shaped to an ellipsoid) to produce the velocity perturbation acting on each focal fish. Surge, sway, and yaw dynamics are governed by Newtonian equations in which hydrodynamic forces and moments are computed from DNS-fitted drag, lift, and moment coefficients, while social interactions enter through a vision-limited attraction and alignment torque with attention parsimony. The wake-induced organization emerges because the modeled oblique jet has narrow regions of favorable forward velocity near its edges, and the superposition of aligned wakes strengthens these edge regions as they propagate downstream.","core_discovery":"The paper's key discovery is that incorporating a phenomenological model of the oblique vortex wake of a carangiform swimmer into a Newtonian agent-based school model produces significantly more organized school topologies than potential-flow-only models. In the high-alignment regime ($\\alpha_T = 0.1$), where attraction is weak relative to alignment, wake-on simulations yield an oblique 'diamond' pattern in which each follower sits near the edge of the wake of the fish ahead; the first principal component of the reconstructed fish distribution explains 11.95% of the variance, compared with 1.54% when the wake is disabled. In the high-attraction regime ($\\alpha_T=0.9$), the wake still improves spatial coherence (3.07% versus 1.42%) but the oblique structure is less pronounced because strong attraction overrides the passive hydrodynamic drafting. The paper further shows that wider wakes (lower Reynolds number or smaller caudal-fin aspect ratio) enhance ordering in high-alignment schools, and that there is an optimal vortex strength beyond which organization degrades.","pith_inferences":["A testable extension would be to apply the same PCA pipeline to experimental tracking of real fish schools; if wake-driven ordering is real, the first-PC explained variance should rise when visual or social cues are experimentally reduced.","The assumption that each fish's wake is unaffected by neighbors is likely the first to fail in dense schools; including wake-wake interactions could shift the predicted optimal vortex strength and should be tested against multi-fish DNS.","The emergence of diamond and staggered patterns offers a hydrodynamic rationale for the inline and phalanx configurations debated in the fish-schooling literature, and could inform bio-inspired design of underwater vehicle formations."],"forward_implications":["Wakes act as an ordering mechanism, so school shape is not set by social rules alone; hydrodynamic history matters.","Fish that generate wider wakes, such as those swimming at lower Reynolds numbers or with smaller caudal-fin aspect ratios, should show more ordered diamond formations when highly polarized.","There is an optimal wake strength: too-strong vortices partially destabilize the school, implying a sweet spot for hydrodynamic schooling benefits.","The model predicts that wake-induced organization is strongest when alignment dominates over attraction, so species or contexts with weak social attraction should exhibit the clearest hydrodynamic patterning."],"supporting_citations":[{"why":"Supplies the DNS-derived wake velocity profile, vortex parameters, and oblique-wake structure used to parameterize the model.","marker":"[13]"},{"why":"Provides the potential-flow-only hydrodynamic fish model that this paper extends and contrasts with the wake-inclusive model.","marker":"[11]"},{"why":"Immersed-boundary flow solver (ViCar3D) used to conduct the DNS that parameterizes hydrodynamic forces and wake features.","marker":"[40]"},{"why":"Bainbridge equation giving tail-beat frequency from swimming speed, used to set the vortex shedding timing.","marker":"[39]"},{"why":"Establishes the oblique vortex pattern of low-aspect-ratio flapping foils that the wake model reproduces.","marker":"[43]"},{"why":"Hypothesis of favorable hydrodynamic regions outside the wake, which the paper's diamond pattern supports.","marker":"[4]"},{"why":"Agent-based model of collective motion with attraction and alignment rules that the social-force component builds on.","marker":"[33]"},{"why":"Data-driven fish school model providing social-force functional forms and the $\\alpha_T$ attraction/alignment weighting parameter.","marker":"[35]"}],"fun_headline_variants":["Hydrodynamic wakes, not social rules, set fish school patterns","Wake vortices order fish schools into diamond patterns","Fish school diamonds emerge from hydrodynamic wakes","Modeling wakes improves fish school organization","Wake strength tunes fish school ordering"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The multi-fish hydrodynamic interaction is represented as a linear superposition of a potential flow and discrete Rankine vortices whose parameters were fitted to a single-fish DNS, and the wake of a fish is assumed not to be modified by trailing or adjacent fish; the paper's conclusion about wake-induced school organization would not hold if this simplified wake representation is inaccurate in multi-fish configurations.","fun_headline_variants_meta":{"raw":{"variants":["Hydrodynamic wakes, not social rules, set fish school patterns","Wake vortices order fish schools into diamond patterns","Fish school diamonds emerge from hydrodynamic wakes","Modeling wakes improves fish school organization","Wake strength tunes fish school ordering"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00037,"raw_usage":{"total_tokens":1987,"prompt_tokens":956,"completion_tokens":1031,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":572,"completion_tokens_details":{"reasoning_tokens":964}},"tokens_in":572,"tokens_out":1031,"duration_ms":9889,"temperature":1.0,"reasoning_tokens":964,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T16:26:28.359113+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Simulate a small school (two or three fish) with the same carangiform kinematics using a fully resolved multi-body DNS and measure the induced velocity field behind the leading fish; if the trailing fish do not preferentially settle near the wake-edge drifting regions, or if the first-PC explained variance does not increase when the wake is present, the model's central organizing mechanism is falsified. More directly, compare the modeled oblique Rankine-vortex wake to the actual wake of a fish swimming in a school: if the wake is substantially modified by neighbors, the superposition assumption breaks.","supporting_citations":[{"cited_title":"Filella, F","cited_arxiv_id":null,"evidence_quote":"Supplies the DNS-derived wake velocity profile, vortex parameters, and oblique-wake structure used to parameterize the model."},{"cited_title":"Zhou, J.-H","cited_arxiv_id":null,"evidence_quote":"Provides the potential-flow-only hydrodynamic fish model that this paper extends and contrasts with the wake-inclusive model."},{"cited_title":"W¨ ohl and S","cited_arxiv_id":null,"evidence_quote":"Immersed-boundary flow solver (ViCar3D) used to conduct the DNS that parameterizes hydrodynamic forces and wake features."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Bainbridge equation giving tail-beat frequency from swimming speed, used to set the vortex shedding timing."},{"cited_title":"Solano, C","cited_arxiv_id":null,"evidence_quote":"Establishes the oblique vortex pattern of low-aspect-ratio flapping foils that the wake model reproduces."},{"cited_title":"Pavlov and A","cited_arxiv_id":null,"evidence_quote":"Hypothesis of favorable hydrodynamic regions outside the wake, which the paper's diamond pattern supports."},{"cited_title":"Czirok, H","cited_arxiv_id":null,"evidence_quote":"Agent-based model of collective motion with attraction and alignment rules that the social-force component builds on."}],"review_version":1}