{"id":"418164f6-bd39-4d0a-981c-cd399d789942","arxiv_id":"2411.19790","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Internal tides in the Mediterranean are widespread, generated mainly at three topographic sites, with semidiurnal waves propagating hundreds of kilometers and diurnal waves trapped near the coast.","lead":"Two high-resolution ocean models reveal that internal tides are generated across the Mediterranean Sea at the Gibraltar Strait, the Sicily Strait and Malta Bank, and the Hellenic Arc, with semidiurnal tides traveling hundreds of kilometers into the Algerian, Tyrrhenian, and Ionian seas. The study offers the first basin-wide map of internal tide energy pathways, useful for improving how ocean models represent tidal mixing in a semi-enclosed sea.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Positive-only summation of conversion term C biases reported basin-integrated internal tide generation totals upward; the magnitude of the omitted negative contribution is unquantified.","rationale":"The reader's weakest assumption is the lack of direct internal-tide validation, which is a broad epistemic limitation. I agree that validation is missing, but the most load-bearing and concretely testable flaw is the positive-only summation, because it directly affects a stated quantitative result and can be corrected with existing model output. The reader did list this issue in the rationale but did not make it the weakest assumption, so agreement is partial. The mapping and propagation claims are supported by two-model consistency and are less affected by this bias; hence the verdict remains conditional rather than being strengthened or weakened.","tokens_in":18905,"tokens_out":4112,"duration_ms":37962,"concrete_test":"Recompute the area-summed and area-weighted-mean C for the entire Mediterranean Sea and for the three regions in Table 3 using all grid points, including negative C values, and also compute the net sum. Report both the positive-only sum and the net sum. If the net sum is substantially smaller than the positive-only sum (for example, less than 70% of it) or if the NEMO/ICON ordering changes, the headline energy totals and the factor-of-two model difference should be revised or re-framed as a positive-only estimate.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The quantitative central claim—2.89 GW (NEMO) and 1.36 GW (ICON) for Mediterranean internal tide generation—rests on a positive-only sum of the conversion term C. In Section 3.1 the authors state that \"only points with positive values are included in the calculation\" because negative C values arise from interference. This is not a harmless filter: C is a signed energy conversion rate, and negative values represent local baroclinic-to-barotropic conversion. By discarding them, the total is systematically inflated, and the paper provides no estimate of the omitted contribution. The reported NEMO/ICON ratio of about 2.1 could be partly an artifact of different spatial patterns of negative C rather than a true difference in generation. Because the abstract and Table 3 present these numbers as basin-integrated generation energy, the energy-budget component of the central claim is not robust as reported.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript presents a basin-wide numerical study of internal tides in the Mediterranean Sea using two independently developed high-resolution ocean models (NEMO v3.6 at 1/24° and ICON-O at R2B9), each providing hourly output for March 2022. The authors compute the barotropic-to-baroclinic energy conversion term C, map baroclinic kinetic energy and tidal beams for the M2 and K1 constituents, and extract internal tide wavelengths from wavenumber spectra that are compared with Sturm-Liouville solutions. The main claims are that internal tides are widespread; that the Gibraltar Strait, Sicily Strait/Malta Bank, and the Hellenic Arc are the primary generation sites; that semidiurnal internal tides propagate hundreds of kilometres; and that the basin-integrated internal tide generation is 2.89 GW in NEMO and 1.36 GW in ICON.","tokens_in":19079,"tokens_out":7016,"duration_ms":60399,"significance":"If the results hold, this is the first basin-wide map of internal tide generation and propagation in the Mediterranean and it identifies the Hellenic Arc as a previously undocumented strong generation region. The paper's use of two independent models with qualitatively consistent spatial patterns is a strength, as is the direct computation of C from diagnosed model fields with no parameter inversion. The comparison of spectral peaks to Sturm-Liouville wavelengths is a useful self-consistency check, and the barotropic-tide assessment against TPXO9 in Appendix B is good practice. However, the quantitative energy totals are not currently robust as reported: the positive-only summation of C and the single-month sample mean the headline numbers should be treated as preliminary rather than as a validated energy budget.","major_comments":[{"comment":"The statement that 'only points with positive values are included in the calculation' means that the values 2.89 GW and 1.36 GW in Table 3 are positive-part sums of the signed conversion term C, not the total barotropic-to-baroclinic conversion. Because negative C represents genuine local conversion from baroclinic to barotropic energy through interference, discarding it systematically inflates the totals and could easily change the NEMO/ICON ratio of about 2.1. The authors should report the net (signed) integral, the separate positive and negative contributions (or at least the omitted negative sum), and a sensitivity test of the filter.","section":"Section 3.1, Table 3"},{"comment":"As printed, Eq. (2) is not correct: the right-hand side, U·∇D + (D+z)·∇U, equals +∇·[(D+z)U], while the left-hand side, −∇(D+z)U, appears to denote the negative of that divergence; moreover, (D+z)·∇U is dimensionally inconsistent as a scalar. If the intended formula is the standard barotropic vertical velocity W = −∇·[(D+z)U], the RHS must be −U·∇D − (D+z)∇·U. Since the sign of C determines which points are counted as 'generation' in the positive-only sum, this diagnostic needs to be corrected and verified.","section":"Section 3.1, Eq. (2)"},{"comment":"All results are drawn from a single month (March 2022), yet the abstract and Table 3 present the generation totals as general Mediterranean values. Given the known seasonal variability of internal tides, the authors should either qualify the totals as March 2022 estimates or provide evidence that this month is representative. In addition, Section 5 explicitly acknowledges the absence of validation against internal-tide observations; given the central role of the quantitative conversion rates, a comparison with available in-situ or altimetric internal-tide data in at least one generation region (e.g., Gibraltar or Sicily Strait) would materially strengthen the claims.","section":"Section 2.2 and Section 5"}],"minor_comments":[{"comment":"Equations (A.1)-(A.5) are garbled in places (e.g., 'q = T ∗ 2', 'j = T ∗ − 1 2 , ..., T ∗ 2'); they should be carefully typeset and checked for mathematical correctness.","section":"Appendix A"},{"comment":"The Sturm-Liouville wavelengths are computed from each model's own N2 and mean depth, so the agreement with the spectral peaks is a self-consistency check rather than an independent theoretical validation; the text should state this more explicitly.","section":"Section 3.3 and Table 4"},{"comment":"The boundaries of the regions used for the area-weighted means in Table 3 are not defined quantitatively; providing the masks or coordinates would improve reproducibility.","section":"Table 3"},{"comment":"There are numerous typographical issues, including 'These works suggest' in Section 1, inconsistent spacing around C, and broken author names in the reference list (e.g., 'M¨ uller'); a careful proofread is needed.","section":"General"}],"recommendation":"major_revision","confidential_remarks":"The manuscript fits the scope of Progress in Oceanography and the qualitative findings are likely of interest. The main risk is that the headline energy totals are presented too strongly given the positive-only summation and the single-month diagnostics; the revision should address these before publication. I do not see citation or novelty concerns."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know two things about this paper. First, it delivers something the literature has lacked: a basin-wide map of internal tide generation and propagation in the Mediterranean, using two independent high-resolution models with qualitatively consistent results. Second, the Hellenic Arc as a major generation site is a real find, and it explains the M2 propagation into the Ionian Sea that earlier regional studies missed.\n\nThe paper does a lot right. The two-model design is the right way to make a first map — different grids, different bathymetry, different tidal forcing, and still the main generation sites and propagation pathways line up. The spectral analysis is careful, they distinguish seiche peaks from internal tide peaks, and the Sturm-Liouville mode calculation is a sensible consistency check even if it is not independent validation. They are also honest about the limitations: the conclusions explicitly call for future validation against satellite and cruise data.\n\nThe soft spots are real but concentrated. The basin-integrated totals in Table 3 — 2.89 GW for NEMO, 1.36 GW for ICON — are not robust as reported. As stated in Section 3.1, they retain only positive values of the conversion term C when summing, discarding negative values that come from interference. That systematically inflates the totals, and the paper gives no estimate of the omitted negative contribution. The NEMO/ICON ratio of roughly two could be partly an artifact of different spatial patterns of negative C. Related issues: the analysis covers one month, there are no uncertainty estimates, and the internal tide fields themselves are not compared to observations — only barotropic tides are checked against TPXO9. None of these sink the qualitative mapping, but they mean the energy budget part of the abstract should be treated as preliminary.\n\nThe reader's stress-test concern is valid, and it lands on the exact passage: the positive-only filter is described but not quantified. That is the single load-bearing flaw, and it is fixable by reporting net conversion, or showing the magnitude of excluded negative values, or framing the numbers as a upper bound.\n\nWho gets value from this? Anyone studying Mediterranean circulation, tidal mixing, or regional internal tide dynamics. It will be cited because it is the first basin-wide reference. It deserves a serious referee — a good editor should send it out, with the expectation that the energy totals get strengthened before publication.","headline":"First basin-wide internal tide mapping for the Mediterranean, with a genuinely new generation site at the Hellenic Arc; worth a serious referee despite energy totals that need a fix.","tokens_in":19586,"tokens_out":1077,"would_cite":true,"duration_ms":12682,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Internal tides are widespread in the Mediterranean Sea, with primary sources at the Gibraltar Strait, the Sicily Strait/Malta Bank, and the Hellenic Arc, and semidiurnal waves travelling hundreds of kilometres.","keywords":["internal tides","Mediterranean Sea","barotropic-to-baroclinic energy conversion","NEMO","ICON-O","semidiurnal internal tides","diurnal internal tides","Hellenic Arc"],"falsifier":"Deploy moored current-and-density arrays for at least a lunar month across the Camarinal Sill, the Malta Bank, and the Hellenic Arc, and compute the observed conversion $C = \\int_{-D}^{\\eta} g\\rho' W \\, dz$; if the basin-wide total comes out far below the 1.36–2.89 GW range, or no coherent semidiurnal beams radiate from the Hellenic Arc into the Ionian Sea, the paper's central map is refuted.","tokens_in":18718,"feed_emoji":"🌊","tokens_out":14625,"duration_ms":119275,"temperature":0.7,"pith_summary":"The paper gives the first basin-wide, model-based census of internal tides in the Mediterranean Sea, a nearly enclosed sea whose weak surface tides have made internal waves easy to overlook. Using two independent high-resolution ocean models, it argues that internal tides are widespread, with the strongest generation at the Gibraltar Strait, the Sicily Strait/Malta Bank, and the Hellenic Arc. From those sites, twice-daily (semidiurnal) internal tides travel hundreds of kilometres into the Algerian, Tyrrhenian, and Ionian seas, while daily (diurnal) internal tides stay trapped against the seabed. A sympathetic reader would care because this map changes where ocean models should put tidal mixing: instead of being confined to a few straits, internal-tide energy is carried into deep basins, and the Hellenic Arc appears as a previously unknown source for the eastern Mediterranean.","feed_headline":"Internal tides are widespread across the Mediterranean, models find","feed_subtitle":"Twice-daily internal waves travel hundreds of kilometres from Gibraltar, Sicily, and the Hellenic Arc into deep basins.","key_machinery":"The load-bearing tool is the barotropic-to-baroclinic energy conversion rate $C = \\int_{-D}^{\\eta} g\\rho' W \\, dz$, where $\\rho'$ is the tidal density perturbation and $W$ the vertical tidal velocity; positive $C$ marks places where the surface tide feeds internal waves, and maps of $C$ define the generation sites. Supporting diagnostics are vertical-mean baroclinic kinetic energy maps at M2 and K1 frequencies, wavenumber spectra along propagation paths, and first-mode wavelengths obtained by solving the Sturm-Liouville eigenvalue problem for vertical modes, which computes theoretical wavelengths from the stratification and separates true internal-tide peaks from Mediterranean seiches at nearby frequencies. Two models with different grids, tidal forcing, bathymetry, and mixing closures are run for the same month, so features common to both are treated as the robust signal.","core_discovery":"On the models' own terms, the discovery is a coherent basin-wide geography of internal tides in a sea often treated as nearly tideless. By computing the conversion of barotropic tidal energy into baroclinic (internal) tidal energy, $C = \\int_{-D}^{\\eta} g\\rho' W \\, dz$, from harmonic analysis of hourly model currents and densities, the paper locates the strongest generation at the Camarinal Sill in the Gibraltar Strait, the Sicily Strait and Malta Bank, and along the Hellenic Arc. The M2 (semidiurnal) internal tide is resolved as beams that leave these sites and cross hundreds of kilometres into the Algerian Sea, the Tyrrhenian Sea, and the Ionian Sea, with first-mode wavelengths of roughly 41–72 km depending on model and region; the K1 (diurnal) internal tide remains bottom-trapped near topography, with generation more widespread in the eastern basin in the NEMO model. Basin-integrated energy conversion totals 2.89 GW in NEMO and 1.36 GW in ICON-O, and the paper attributes the difference chiefly to diurnal-tide generation and to differences in bathymetry, stratification, and barotropic tides between the two configurations.","pith_inferences":["The paper does not analyse seasons other than March 2022; a natural extension is that internal-tide generation at the Hellenic Arc and propagation into the Ionian Sea should strengthen or weaken with the seasonal pycnocline, so an annual model run would show whether the mapped geography is permanent.","The gap between the two models' basin totals (1.36 GW vs 2.89 GW) means the true conversion value is bracketed but not pinned down; direct estimates from moored arrays could narrow this range and test whether the Mediterranean's internal-tide energy budget is closer to the lower or upper estimate.","Because neither model resolves sub-kilometre topography such as the Messina Strait, the paper's own list of unresolved sites implies the basin totals may be lower bounds if such narrow straits also generate internal tides."],"forward_implications":["Basin-scale tidal mixing in the Mediterranean should not be treated as local to straits: semidiurnal internal tides carry energy hundreds of kilometres into the Algerian, Tyrrhenian, and Ionian seas before dissipating.","The Hellenic Arc, if confirmed by observations, is a new source region for the eastern Mediterranean's internal wave field and helps explain previously reported semidiurnal energy in the Ionian Sea and Cretan Passage.","The Mediterranean's total internal-tide generation, 1.36–2.89 GW, is small next to the roughly 1.7 TW global conversion, but it falls in a basin where tidal mixing is often set to near zero in regional models, so omitting internal tides would under-mix deep water.","The resolved first-mode M2 wavelengths of roughly 41–72 km are consistent with the models' horizontal resolution, meaning the main beams are captured while higher modes, which dissipate locally, are not."],"supporting_citations":[{"why":"Prior observations of internal tide oscillations at the Camarinal Sill; defines the Gibraltar Strait as a known generation site that the paper's conversion maps confirm.","marker":"Morozov et al. (2002)"},{"why":"Current measurements showing internal tides in the Strait of Sicily; the central Mediterranean generation site is anchored to this earlier evidence.","marker":"Gasparini et al. (2004)"},{"why":"Observational and numerical study of internal tides in the central Mediterranean and Malta Bank; the paper extends this regional result and compares K1 behaviour.","marker":"Oddo et al. (2023)"},{"why":"Cruise measurements identifying propagating semidiurnal internal tides in the Aegean; used as a comparison point for mapped propagation regions.","marker":"Alford et al. (2012)"},{"why":"Supplies the global internal-tide dissipation context and the tide model used to validate the models' barotropic tides against observations.","marker":"Egbert and Ray (2003)"},{"why":"Provides the conversion diagnostic $C$ and the global conversion value of about 1.7 TW against which the Mediterranean totals are compared.","marker":"Müller (2013)"},{"why":"Gives global M2 internal-tide first-mode wavelength ranges and the modelling approach used to identify modes in the wavenumber spectra.","marker":"Li et al. (2015)"},{"why":"Harmonic tidal analysis method used to isolate tidal currents and to derive the conversion and kinetic-energy fields throughout the paper.","marker":"Foreman et al. (2009)"}],"fun_headline_variants":["Mediterranean internal tides traced from Gibraltar to deep basins","Internal tides travel hundreds of km across Mediterranean, models show","High-res models map Mediterranean internal tide hotspots","Semidiurnal internal tides cross Mediterranean for hundreds of km","Gibraltar, Sicily, Hellenic Arc: where Mediterranean internal tides begin"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the two hydrostatic models, validated only for surface tides against a global tide atlas and not against observed internal wave fields, still reproduce where and how strongly internal tides are generated and how far they travel; the paper's own conclusion asks for satellite and cruise data to test this.","fun_headline_variants_meta":{"raw":{"variants":["Mediterranean internal tides traced from Gibraltar to deep basins","Internal tides travel hundreds of km across Mediterranean, models show","High-res models map Mediterranean internal tide hotspots","Semidiurnal internal tides cross Mediterranean for hundreds of km","Gibraltar, Sicily, Hellenic Arc: where Mediterranean internal tides begin"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000803,"raw_usage":{"total_tokens":3574,"prompt_tokens":1038,"completion_tokens":2536,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":654,"completion_tokens_details":{"reasoning_tokens":2455}},"tokens_in":654,"tokens_out":2536,"duration_ms":15893,"temperature":1.0,"reasoning_tokens":2455,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T05:48:46.254084+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Deploy moored current-and-density arrays for at least a lunar month across the Camarinal Sill, the Malta Bank, and the Hellenic Arc, and compute the observed conversion $C = \\int_{-D}^{\\eta} g\\rho' W \\, dz$; if the basin-wide total comes out far below the 1.36–2.89 GW range, or no coherent semidiurnal beams radiate from the Hellenic Arc into the Ionian Sea, the paper's central map is refuted.","supporting_citations":[{"cited_title":", author Smeed, D.A","cited_arxiv_id":null,"evidence_quote":"Current measurements showing internal tides in the Strait of Sicily; the central Mediterranean generation site is anchored to this earlier evidence."},{"cited_title":", author Poulain, P","cited_arxiv_id":null,"evidence_quote":"Observational and numerical study of internal tides in the central Mediterranean and Malta Bank; the paper extends this regional result and compares K1 behaviour."},{"cited_title":", author Gregg, M.C","cited_arxiv_id":null,"evidence_quote":"Cruise measurements identifying propagating semidiurnal internal tides in the Aegean; used as a comparison point for mapped propagation regions."},{"cited_title":", author Ray, R.D","cited_arxiv_id":null,"evidence_quote":"Supplies the global internal-tide dissipation context and the tide model used to validate the models' barotropic tides against observations."},{"cited_title":", author Cherniawsky, J.Y","cited_arxiv_id":null,"evidence_quote":"Harmonic tidal analysis method used to isolate tidal currents and to derive the conversion and kinetic-energy fields throughout the paper."}],"review_version":1}