{"id":"276f5c6d-6761-4ce5-99a2-008f18fdc57b","arxiv_id":"2506.19081","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"An agent-based model shows that swimming persistence and the ratio of downward to upward swimming speed organize plankton diel vertical migration into four regimes.","lead":"The authors simulate swarms of light-sensitive swimmers and find four collective patterns, from surface crowding to permanent sinking, controlled by two simple numbers: how persistently each swimmer holds its course and how much faster it swims down than up. The framework aims to explain why real plankton migrations vary so widely, and could help predict how migrating organisms respond to changing light environments.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Two-parameter universality is asserted rather than demonstrated: the (Pe,W) phase map comes from a single isolume threshold and one fixed set of behavioral parameters, with no sensitivity test.","rationale":"The reader's weakest_assumption identifies the same load-bearing concern I would put first: the phase diagram is computed on a 2D slice of a higher-dimensional parameter space, and the paper provides no sensitivity analysis for the fixed choices. My reading of the main text, Methods, and Fig. 5 confirms that only D_R and w_down are varied, while I*, w_up, tau_G, H, L, R, k_d, and the day-night timescale are held constant. The phrase 'without loss of generality' attached to I* = 0.85 I0 is doing substantial logical work, and it is not backed by an argument or a control. The concrete I* test would settle the most direct version of this objection. In addition, I checked the analytical equations and found an internal inconsistency: Eq. 4 is not the W-scaled form of Eq. 3. This does not by itself invalidate the simulated regime classification, but it does undermine the analytical estimates of regime boundaries, including Eq. 5, which is asserted without derivation. The paper has genuine strengths: it is an agent-based model with explicit Langevin dynamics, WCA interactions, and a reported 120-run parameter sweep, and the four regimes are described with concrete statistical observables. Those features make the classification plausible as a simulation result. The unresolved issue is the universality claim and the unverified analytical scaffolding, not the existence of the regimes. Conditional acceptance is therefore the right verdict, and it matches the reader's CONDITIONAL decision, so I recommend no change to the verdict.","tokens_in":15198,"tokens_out":8075,"duration_ms":86041,"concrete_test":"Repeat the full (Pe,W) sweep of Fig. 5 with the isolume threshold changed to I* = 0.50 I0, keeping all other parameters (L, R, w_up, tau_G, H, initialization, and integration settings) unchanged, and overlay the regime boundaries obtained from the mean-position and kurtosis maps. If the boundaries move by more than the native resolution of the sampled grid (e.g., more than half a decade in Pe or 0.25 in W), the 'without loss of generality' claim for I* fails and the two-parameter description is not sufficient as stated.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that Pe and W alone govern the four collective regimes is load-bearing and under-supported. All 120 simulations fix I* = 0.85 I0, w_up = 2 u0, tau_G = 1, the niche depth H, and the response parameters L and R; only D_R and w_down are varied. The text calls the isolume choice 'without loss of generality,' but no scaling argument or control simulation supports that. The swimming rule in Eq. 3 uses the absolute values of w_up and w_down inside the tanh switches, and the function S(z,t) determines how quickly the photokinetic response saturates, so W alone does not control the response unless all other speeds and light-response scales are fixed. A second, independent issue weakens the analytical framing: Eq. 4 is not consistent with Eq. 3. Dividing Eq. 3 by w_up for I < I* gives w/w_up = 1/2 + (1/2)tanh(A1) + W tanh(A2) - W/2, whereas Eq. 4 is 1/2 + (1/2)(tanh A1 + tanh A2) - W/2; the difference is (W - 1/2)tanh(A2). Eq. 5 is also presented without derivation. The four regimes may well be robust within the tested slice, but that slice is a 2D section of a higher-dimensional parameter space. The abstract's 'governed by just two dimensionless parameters' therefore overstates what the simulations establish.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript presents a three-dimensional agent-based model of self-propelled photokinetic swimmers that respond to a dynamic light field through a vertical speed modulation rule (Eq. 3), combined with gyrotactic reorientation and rotational noise. The authors perform 120 Brownian dynamics simulations varying the rotational diffusion coefficient and the downward swimming speed, identify four collective regimes (Surface Accumulation, Shallow DVM, Deep DVM, and Sinking), and propose that population-scale dynamics are governed by just two dimensionless parameters: the Péclet number Pe and the vertical speed ratio W. The regimes are characterized through the mean and excess kurtosis of the time- and horizontally-averaged vertical concentration profile, and analytical boundary conditions are proposed in Eqs. (4) and (5). The paper argues that a feedback between upward-biased swimming and gyrotactic alignment drives surface accumulation, while excessive downward bias leads to irreversible sinking.","tokens_in":15510,"tokens_out":7107,"duration_ms":67996,"significance":"If substantiated, the proposed two-parameter organization of diel vertical migration regimes would provide a valuable mechanistic bridge from individual photokinetic responses to population-scale patterns, with potential implications for aquatic ecology and active-matter physics. The simulation study is systematic within its chosen parameter slice: the four regimes are visually and statistically distinguishable in the examples shown, and the use of distribution shape descriptors (mean and excess kurtosis) is a reasonable and transparent way to classify collective behavior. The inclusion of hard-core interactions and stochastic reorientation strengthens the model's realism. However, the central universality claim rests on a single set of fixed behavioral parameters, and the analytical formulas supporting the phase diagram contain algebraic and conceptual inconsistencies. The qualitative regime classification is plausible, but the quantitative claim of two-parameter control is not yet established.","major_comments":[{"comment":"Eq. (4) does not follow from Eq. (3) for general W. For the branch I < I*, Eq. (3) gives w/w_up = 1/2 + (1/2)tanh(A1) + W tanh(A2) - W/2, whereas Eq. (4) reads 1/2 + (1/2)(tanh(A1) + tanh(A2)) - W/2. The two expressions differ by (W - 1/2)tanh(A2), which vanishes only for W = 1/2 or tanh(A2) = 0. Since Fig. 5d and the subsequent feedback argument are built directly on Eq. (4), this algebraic discrepancy is load-bearing and must be corrected or the derivation must be revised.","section":"Mapping and bounding collective patterns, Eq. (4)"},{"comment":"Eq. (5) is introduced as a hypothesis without a derivation linking the zero-vertical-velocity condition to the gyrotactic orientation average <p> = coth(λ) - 1/λ. In addition, substituting the stated value w_up = 2u0 into Eq. (5) gives W = 1 + coth(λ) - 1/(2λ), which is always greater than about 1.8 and does not approach 1 as λ grows; this contradicts the following sentence claiming that the smallest W satisfying the condition is 1. The authors need to provide a step-by-step derivation of Eq. (5) and reconcile its predictions with the parameter values used in the simulations.","section":"Mapping and bounding collective patterns, Eq. (5)"},{"comment":"The central claim that the four regimes are governed solely by Pe and W is not supported by the evidence as presented, because all 120 simulations fix I* = 0.85 I0, w_up = 2u0, τ_G = 1, the niche depth H, the response parameters L and R, and the day-night light shape parameters. The phrase 'without loss of generality' is not justified by any scaling argument or control simulations. As written, the (Pe, W) phase diagram is a two-dimensional slice of a higher-dimensional parameter space. The authors should provide sensitivity analyses showing that the regime boundaries are insensitive to these fixed choices, or explicitly revise the claims to describe the behavior within this slice.","section":"Population-scale collective patterns and Fig. 5"},{"comment":"The phase diagram lacks uncertainty quantification and a precise description of how the colored boundary dots are determined. Since the regime labels are defined by threshold values of μ/L_z and α, the reader cannot assess the robustness of the boundaries to statistical fluctuations, finite simulation time, or alternative threshold choices. Please specify the classification algorithm and include error bars or a sensitivity analysis of the thresholds.","section":"Mapping and bounding collective patterns, Fig. 5(a–b)"}],"minor_comments":[{"comment":"The symbol W is used both for the vertical swimming speed field in Eq. (6) and for the dimensionless ratio w_down/w_up throughout the rest of the paper; please rename one of them to avoid confusion.","section":"Methods, Eq. (6)"},{"comment":"Please clarify how the noise term in Eq. (7) relates to the rotational diffusion coefficient D_R. With the prefactor 1/γ_R and the stated correlation ⟨ξ_R(t1)ξ_R(t2)⟩ = D_R δ(t1-t2), the effective angular diffusivity is D_R/γ_R^2 unless γ_R = 1; the paper should state the value of γ_R used and how the numerical noise amplitude was chosen.","section":"Methods, Eq. (7)"},{"comment":"The function S(z,t) is defined as S(z,t) = ∂I(z,t)/∂t + k_d, which mixes units of I/time and 1/length; please provide the full nondimensionalization and verify the expression, as this function controls the tanh switches in Eq. (3).","section":"ABM modeling framework, Eq. (3) and definitions"},{"comment":"The caption for Fig. 4 has a typographical error: 'Deep DVM (Pe=10, W=0.8' is missing a closing parenthesis; it should read '(Pe=10, W=0.8)'.","section":"Fig. 4 caption"},{"comment":"The sentence beginning 'This behavior has been observed in phytoplankton communities like H. akashiwo [20]' relies on a reference about sharp turns and gyrotaxis in a different context; please explain explicitly how the cited experiments support the 'floating' mechanism described.","section":"Discussion, deep DVM paragraph"},{"comment":"The statement that the model is the 'first fully three-dimensional agent-based framework that reproduces canonical DVM behavior' may overstate novelty given the existing agent-based DVM literature; consider softening the claim or providing a more precise comparison.","section":"Discussion, deep DVM paragraph"}],"recommendation":"major_revision","confidential_remarks":"The paper's qualitative regime classification is plausible and the simulation effort is substantial, but the two-parameter universality claim is broader than the evidence. The algebraic error in Eq. (4) and the inconsistency in Eq. (5) need to be fixed, and a sensitivity analysis with respect to the fixed behavioral parameters is essential before the central claim can be evaluated. If the authors can provide these, the manuscript may become suitable for publication. I would also encourage the authors to make the simulation code available, since the data availability statement mentions only the figure data."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This is a genuinely new simulation result: a four-regime phase diagram for a photokinetic agent-based model with gyrotaxis and rotational noise, and the four regimes are clearly visible in the output. The soft spot is that the paper sells a \"two-parameter\" story without testing whether the other fixed parameters actually matter, and the one analytical equation meant to explain the feedback doesn't follow from Eq. 3.\n\nWhat's new and good: the model is a clean, physically motivated assembly of Richards et al.'s photokinetic swimming rule and Pedley-Kessler gyrotaxis in 3D. The single-particle comparison (Brownian vs light-driven) is a nice illustration of how a day-night light field converts bounded stochastic exploration into a much larger vertical excursion. The phase diagrams based on mean position and excess kurtosis are a reasonable way to organize the phenomenology, and the qualitative links to real organisms (Daphnia, Akashiwo, thin layers) are appropriate. The four regimes—Surface Accumulation, Shallow DVM, Deep DVM, Sinking—do emerge naturally from the simulations, and the \"sinking\" collapse at W>1 is a simple and intuitive threshold.\n\nSoft spots: First, Eq. 4. Dividing Eq. 3 by w_up gives 1/2 + (1/2)tanh(A1) + W tanh(A2) - W/2, not Eq. 4's 1/2 + (1/2)tanh(A1) + (1/2)tanh(A2) - W/2. The error is (W - 1/2)tanh(A2). So the \"positive feedback\" discussion and anything built on Eq. 4 is not a correct consequence of the swimming rule. Second, Eq. 5 is presented as a hypothesis with no derivation. It might be a reasonable curve, but the reader cannot check it. Third—and most important for the central claim—the phase diagram uses one isolume threshold (I*=0.85 I0), one upward speed (w_up=2 u0), one gyrotactic timescale (tau_G=1), and fixed light-response parameters. The text calls the isolume choice \"without loss of generality,\" but no scaling or control simulation is given. The swimming strategy in Eq. 3 uses absolute values of w_down and w_up inside tanh switches, so W alone cannot determine the response unless those speeds and light-response scales are fixed. The 120 simulations vary only Pe and W; that's a 2D slice of a higher-dimensional space. The abstract's \"governed by just two dimensionless parameters\" therefore overstates what is actually demonstrated.\n\nThese are fixable issues, but they need to be addressed before the analytical framing is taken seriously. The paper will interest people working on phototactic active matter and on mechanistic models of DVM. It deserves a serious referee; the simulation result is plausible and worth airing. The referee should insist on sensitivity tests and a corrected, clearly labeled heuristic derivation.","headline":"A plausible and useful four-regime phase diagram for photokinetic swimmers, undermined by an overstated two-parameter universality claim and an inconsistent analytical aside.","tokens_in":16069,"tokens_out":3600,"would_cite":true,"duration_ms":33891,"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":"An agent-based model reduces diel vertical migration to two dimensionless parameters, mapping four collective regimes.","keywords":["diel vertical migration","agent-based model","photokinesis","gyrotaxis","Péclet number","collective behavior","active matter","light niche"],"falsifier":"Repeat the $(Pe,W)$ scan with the isolume threshold changed to, say, $I^*=0.5I_0$ or $0.95I_0$, or with $w_{\\mathrm{up}}$ varied, and check whether the locations of the four regime boundaries in the $(Pe,W)$ plane shift by more than a small tolerance; if they do, the claimed two-parameter control fails. Alternatively, measure $w_{\\mathrm{down}}/w_{\\mathrm{up}}$ and persistence length for a species such as Daphnia or Heterosigma akashiwo and test whether the predicted Deep-DVM band $0.3\\lesssim W\\lesssim 1.25$ matches observed vertical distributions.","tokens_in":14987,"feed_emoji":"🌊","tokens_out":4538,"duration_ms":46314,"temperature":0.7,"pith_summary":"The paper claims that the large-scale collective patterns of light-driven swimming organisms, notably the daily vertical migration of plankton, can be captured by an agent-based model in which each swimmer follows simple local rules: self-propulsion, gyrotactic reorientation, rotational noise, and a vertical speed that responds to local light relative to a preferred isolume. Solving the Langevin dynamics of a swarm produces four robust population-level regimes, Surface Accumulation, Shallow DVM, Deep DVM, and Sinking, organized by just two dimensionless parameters: the Péclet number $Pe$ (persistence versus rotational diffusion) and the vertical speed ratio $W=w_{\\mathrm{down}}/w_{\\mathrm{up}}$ (photokinetic asymmetry). The central insight is a feedback: upward-biased swimming reinforces surface aggregation, while sufficient downward bias is irreversible and collapses migration. If correct, this offers a mechanistic bridge from organism-scale behavior to the statistical vertical distributions observed in lakes and oceans.","feed_headline":"Two numbers govern how plankton swarms migrate","feed_subtitle":"Persistence and vertical speed ratio predict four collective regimes, from surface buildup to irreversible sinking.","key_machinery":"The central object is the coupled Langevin system for each swimmer, Eqs. (6)-(7): position updates from self-propulsion $u_0\\hat{p}_i$, repulsive collisions, and a light-modulated vertical swimming speed $W(z,t)$, plus orientation updates from rotational white noise and a gyrotactic torque that reorients the body against gravity on timescale $\\tau_G$. The light-dependent swimming strategy, Eq. (3), compares local intensity to the isolume $I^*=0.85I_0$ and sets upward and downward speeds through tanh response functions, so the two parameters $Pe=u_0/(D_R\\sigma)$ and $W=w_{\\mathrm{down}}/w_{\\mathrm{up}}$ enter through the noise amplitude and the speed asymmetry. The analytical boundary estimate, Eq. (5), comes from setting the mean vertical velocity to zero at the isolume, using the Fokker-Planck polarization $\\langle p\\rangle=\\coth(\\lambda)-1/\\lambda$ with $\\lambda=Pe[\\sigma/(\\tau_G u_0)]$.","core_discovery":"The discovery, on the paper's own terms, is that macroscopic diel vertical migration is not a single phenomenon but a family of four collective regimes, and the regime a population lands in is set almost entirely by two numbers. In simulations spanning $Pe$ and $W$, a sharply peaked distribution at the nocturnal niche marks Shallow DVM, a peaked distribution with a deep tail synchronized with the moving isolume marks Deep DVM, confinement near the surface marks Surface Accumulation, and a broad deep distribution marks irreversible Sinking. The paper shows that Deep DVM, the canonical migration, requires balanced upward and downward speeds with $W\\approx 1$, and that the transition to Sinking can be estimated analytically from the condition that the mean vertical velocity vanishes, giving a critical $W$ that is never below 1. It also reports that when $Pe<1$, where stochastic reorientation dominates, the collective pattern is largely insensitive to $Pe$ and shaped mainly by $W$.","pith_inferences":["If the two-parameter collapse is robust to the fixed isolume threshold, then field data could be inverted: acoustic backscatter profiles give mean depth and kurtosis, which map onto $(Pe,W)$, yielding organism-level behavioral estimates from population-level observations. (Beyond the paper.)","The model omits fluid advection and turbulence; since real DVM occurs in stratified, sheared columns, adding turbulence would effectively renormalize $Pe$ and could shift the delicate $W\\approx 1$ Deep-DVM balance, a concrete testable extension.","The prediction that $W=1$ is the lower bound for sinking regardless of $\\tau_G$ suggests a generic design rule for light-responsive active matter: any mechanism biasing vertical motion downward beyond equality with upward speed will drive irreversible loss from the illuminated zone, testable with synthetic phototactic colloids.","Changing water turbidity alters the isolume depth and thus the effective light gradient; the paper's emphasis on gradients rather than absolute intensity implies that turbidity could shift regime boundaries without changing organism-level swimming parameters, which field comparisons could check. (Beyond the paper.)"],"forward_implications":["A population with balanced upward and downward swimming speeds ($W\\approx 1$) should show canonical deep DVM across a broad range of Péclet numbers, whereas strong upward bias ($W\\ll 1$) confines the swarm to the upper water column.","Once $W$ exceeds the critical value derived from the $w=0$ condition, migration collapses into irreversible Sinking for all $Pe$, so a modest increase in downward bias can eliminate DVM entirely.","In noise-dominated populations ($Pe<1$), the regime is controlled almost entirely by the vertical speed ratio $W$, not by persistence, so behavioral asymmetry rather than swimming strength dictates the population pattern.","The regime boundaries can be read off time-averaged vertical distributions through the mean position and excess kurtosis, the same statistics that acoustic backscatter profiles provide.","The minimum critical $W$ for the Sinking transition equals 1 independently of the gyrotactic timescale, giving a simple, testable threshold for loss of migration."],"supporting_citations":[{"why":"Supplies the photokinetic swimming-strategy model, Eq. (3), that sets vertical velocity from light intensity relative to the isolume.","marker":"[31]"},{"why":"Provides the Fokker-Planck result for gyrotactic orientation polarization used in the analytical boundary estimate.","marker":"[42]"},{"why":"Supplies the continuum gyrotactic modeling framework and the gravitational torque term used in the orientation dynamics.","marker":"[46]"},{"why":"Supports the gyrotactic transport equations for spherical swimmers used to model orientation in the Langevin system.","marker":"[58]"},{"why":"Provides the empirical example of sharp turns and gyrotaxis modulating surface accumulation, used to interpret the Deep DVM and reorientation mechanisms.","marker":"[20]"},{"why":"Supplies global acoustic estimates of DVM migration times and velocities, supporting the parameter ranges and the connection to observed migration speeds.","marker":"[57]"},{"why":"Supplies the one-dimensional ocean-model context for DVM and the biological-pump implications that motivate the population-scale question.","marker":"[2]"}],"fun_headline_variants":["Two numbers decide if plankton swim, sink, or stay","Plankton migration: a single ratio can tip a swarm into sinking","Balance the speeds: deep migration needs near-equal up and down","Four plankton regimes traced to two dimensionless parameters","Surface or deep: the photokinetic ratio that shapes DVM"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that fixing the isolume threshold at $I^*=0.85I_0$, the upward speed $w_{\\mathrm{up}}=2u_0$, the gyrotactic timescale $\\tau_G=1$, and the response parameters $L$ and $R$ does not change the qualitative phase boundaries, so that $Pe$ and $W$ alone govern the regimes.","fun_headline_variants_meta":{"raw":{"variants":["Two numbers decide if plankton swim, sink, or stay","Plankton migration: a single ratio can tip a swarm into sinking","Balance the speeds: deep migration needs near-equal up and down","Four plankton regimes traced to two dimensionless parameters","Surface or deep: the photokinetic ratio that shapes DVM"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00108,"raw_usage":{"total_tokens":4515,"prompt_tokens":941,"completion_tokens":3574,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":557,"completion_tokens_details":{"reasoning_tokens":3490}},"tokens_in":557,"tokens_out":3574,"duration_ms":26233,"temperature":1.0,"reasoning_tokens":3490,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T18:37:21.940138+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Repeat the $(Pe,W)$ scan with the isolume threshold changed to, say, $I^*=0.5I_0$ or $0.95I_0$, or with $w_{\\mathrm{up}}$ varied, and check whether the locations of the four regime boundaries in the $(Pe,W)$ plane shift by more than a small tolerance; if they do, the claimed two-parameter control fails. Alternatively, measure $w_{\\mathrm{down}}/w_{\\mathrm{up}}$ and persistence length for a species such as Daphnia or Heterosigma akashiwo and test whether the predicted Deep-DVM band $0.3\\lesssim W\\lesssim 1.25$ matches observed vertical distributions.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the photokinetic swimming-strategy model, Eq. (3), that sets vertical velocity from light intensity relative to the isolume."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the Fokker-Planck result for gyrotactic orientation polarization used in the analytical boundary estimate."},{"cited_title":"Pedley and J","cited_arxiv_id":null,"evidence_quote":"Supplies the continuum gyrotactic modeling framework and the gravitational torque term used in the orientation dynamics."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supports the gyrotactic transport equations for spherical swimmers used to model orientation in the Langevin system."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the empirical example of sharp turns and gyrotaxis modulating surface accumulation, used to interpret the Deep DVM and reorientation mechanisms."},{"cited_title":"Bianchi and K","cited_arxiv_id":null,"evidence_quote":"Supplies global acoustic estimates of DVM migration times and velocities, supporting the parameter ranges and the connection to observed migration speeds."}],"review_version":2}