{"id":"453f9b4c-ccc2-4441-bc5c-5beb95232788","arxiv_id":"2412.10601","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":7,"one_line_summary":"A new inversion scheme combines ground observatory and multi-satellite data in three-hour windows to produce continuous 2014-2023 time series of ionospheric, magnetospheric, and induced magnetic field coefficients without temporal harmonic priors.","lead":"Researchers built a new model that separates magnetic signals from the ionosphere, the magnetosphere, and the Earth's interior using a decade of ground and satellite measurements. The model runs in short time windows without assuming fixed daily or seasonal cycles, making it useful for studying geomagnetic storms and for probing the Earth's interior conductivity.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The thin-sheet ionosphere relation (Eq. 12) is asserted but never validated, so the central ground/satellite separation claim depends on an untested structural assumption that could bias all three source estimates.","rationale":"I agree with the reader's identification of the weakest assumption. The thin-sheet relation of Eq. (12) is the mathematical linchpin of the source separation in Eqs. (13)-(14). The paper never specifies h, never quantifies the error from the thin-sheet approximation, and never tests whether the data themselves support the relation between qion and gion. Because the ionosphere is a finite-thickness layer with field-aligned currents, especially during storms, the assumed one-to-one relation is physically approximate rather than exact. Any violation propagates directly into the estimated induced and magnetospheric coefficients, which is precisely the separation the paper claims. The paper has real independent support: the 5-fold cross-validation shows the ionosphere improves generalization, the magnetospheric qmag_10 agrees with CI and CHAOS during quiet times and storms, and lunar tidal peaks appear only in ionospheric coefficients while being absent in magnetospheric ones. Those checks are strong evidence that the separation is capturing physical signals, but they do not validate the thin-sheet relationship itself. A free-ionosphere inversion, or equivalently varying h over the plausible E-region range, would directly test whether the separation is robust to the structural assumption. Since the concern is addressable and does not by itself refute the empirical results, the appropriate verdict remains CONDITIONAL, matching the reader's assessment.","tokens_in":37802,"tokens_out":19053,"duration_ms":201447,"concrete_test":"Re-run the per-bin inversion on the same dataset with a free ionosphere: replace the satellite internal coefficients in Eq. (14) with gint + gion, estimating gion independently (same truncation as qion), while keeping ground external qion+qmag and satellite external qmag common. Compare the recovered qmag_10, qmag_21, gint_10, and gint_21 time series against the thin-sheet-constrained results, and check whether the free estimates satisfy gion ≈ -[n/(n+1)]((a+h)/a)^(2n+1) qion within bootstrap uncertainties. If the coefficient time series shift by more than the reported RMS differences (e.g., 0.86-2.4 nT for qmag_10), the thin-sheet assumption is not supported by the data and the separation is model-dependent.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim of tractable three-way separation is implemented through Eqs. (13)-(14), where the satellite internal field is written as gint - [n/(n+1)]((a+h)/a)^(2n+1) qion. This substitution is the only mechanism that connects the ionospheric field seen by ground observatories (external coefficients qion) to the ionospheric field seen by satellites (internal coefficients gion). It rests entirely on the thin-sheet continuity condition of Eq. (12). The manuscript nowhere states the numerical value of h, and it provides no test of whether the imposed relation is consistent with the actual data. Real ionospheric currents have finite vertical extent and field-aligned components, particularly during the geomagnetic storms that this paper explicitly targets. If Eq. (12) is violated, the inversion will trade off the misfit between ground and satellite ionospheric signals against gint and qmag, leaking ionospheric signal into the induced and magnetospheric coefficient time series. The paper's internal consistency checks (no-ionosphere comparison, lunar peaks appearing only in ionospheric coefficients) demonstrate that the ionosphere matters, but they do not establish that the specific thin-sheet parametrization is correct. Because Eqs. (13)-(14) are the heart of the claimed separation, an unvalidated Eq. (12) is the most load-bearing weakness of the paper.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a new geomagnetic field modeling approach that simultaneously estimates mid-latitude ionospheric, magnetospheric, and internally induced magnetic fields by combining ground observatory and multi-satellite (Swarm, CryoSat-2, Grace-FO) vector data over 2014-2023. The method works in short (3-hour) time bins, imposes no temporal harmonic structure, and uses a thin-sheet approximation for ionospheric currents to link ground and satellite observations. Coefficients are estimated by robust least squares with model complexity selected via cross-validation. The resulting 10-year time series are analyzed in the frequency domain, revealing solar, lunar, and storm-time signatures, and are used to compute electromagnetic transfer functions (C-responses) that are more coherent and physically plausible when the ionosphere is explicitly modeled. The paper also demonstrates that omitting the ionosphere significantly biases induced-field estimates, especially at higher degrees.","tokens_in":38011,"tokens_out":3162,"duration_ms":32713,"significance":"If the separation is valid, this is a substantial methodological advance: it enables continuous, all-local-time, all-magnetic-condition monitoring of external and induced fields without prescribing temporal harmonics, which is relevant for space weather nowcasting and for electromagnetic induction studies. The paper provides several strong cross-checks: lunar tidal peaks appear only in ionospheric coefficients, the Sq current vortices are recovered with expected seasonal behavior, magnetospheric coefficients agree with CI and CHAOS models for the dominant modes, and the C-responses computed with the ionosphere included are smoother and more coherent. The availability of the coefficient time series on Zenodo and the use of cross-validation for model selection are also positive features. The main risk is that the central separation rests on an unquantified and unvalidated thin-sheet assumption for the ionosphere, and that no formal uncertainties accompany the coefficient time series.","major_comments":[{"comment":"The thin-sheet relation (Eq. 12) is the only mechanism that connects the ionospheric field as seen by ground observatories (qion) and by satellites (gion), and it is therefore load-bearing for the central three-way separation claim. The manuscript does not state the numerical value of the sheet height h, nor does it provide any validation that the imposed radial Br continuity holds for the real ionosphere during the geomagnetic storms that the paper explicitly targets. Real ionospheric currents have finite vertical extent, field-aligned components, and day-night conductivity asymmetries, all of which violate Eq. (12) to some degree. I request that the authors: (i) specify h and justify the choice; (ii) perform a sensitivity analysis varying h over a plausible range (e.g., 90-120 km) and report how the separated coefficients change; and (iii) assess the error introduced by the thin-sheet approximation, for example by comparing against a model with a vertically extended ionospheric current layer or against independent ionospheric field estimates. Without such tests, the possibility that ionospheric signal leaks into the magnetospheric and induced coefficients through the parametrization in Eqs. (13)-(14) cannot be ruled out.","section":"Section 2, Eq. (12)"},{"comment":"The paper reports numerous spectral peaks (e.g., at 29.7, 27.0, 25.5 days; Rieger-type periods; lunar tidal lines) as robust features of the reconstructed fields, but no formal uncertainties are given for the estimated Gauss coefficient time series. Given that the separation depends on a structural assumption and that the data coverage changes over time (Fig. 3), it is important to know whether the claimed peaks are statistically significant. I ask the authors to provide at least approximate uncertainties, for example via bootstrap resampling of the time bins, jackknife estimates, or posterior covariances from the least-squares problem in each time bin, and to indicate the impact on the spectral interpretation.","section":"Section 5.3 (and throughout)"},{"comment":"The observatory biases are estimated from a model that omits the ionosphere, under the argument that any static offset is small and does not affect temporal variability. However, the bias estimation is a separate regression on quiet-night data, and the resulting biases are then subtracted from the full dataset used in the main inversion. The sensitivity of the final separated coefficients to this preprocessing step is not quantified. Please provide a comparison of results obtained with the CI-derived biases versus the in-house biases, or a perturbation test that adds a plausible static offset to the biases and shows that the ionospheric, magnetospheric, and induced coefficient time series and the derived C-responses are materially unchanged.","section":"Section 3.2.1 and Appendix B"}],"minor_comments":[{"comment":"The radical notation in Eq. (17) and the formatting of Eq. (20) are inconsistent and difficult to parse; please rewrite them in standard LaTeX style.","section":"Equations (17) and (20)"},{"comment":"There is a typo: \"spacial\" should be \"spatial\" in the sentence beginning \"To get a glimpse of the spacial structure\" (Section 6, first paragraph).","section":"Section 5.2, text near Figs. 7-10"},{"comment":"The text contains several placeholder references such as \"Fig. ??\", \"Table ??\", and \"Fig. ??\" (e.g., Section 3.2.1 and Section 5.3). These need to be resolved to the actual supplementary figure and table numbers.","section":"Supplementary figures"},{"comment":"The label \"Kp 2\" in the figure panels is ambiguous; it should read \"Kp ≤ 2\" for clarity, matching the text.","section":"Figures 4 and 5"},{"comment":"The statement \"The uncertainty reported here is the resolution at the given peaks\" applies to some but not all listed peak periods; please state the frequency resolution explicitly and apply it consistently to all quoted periods.","section":"Section 5.3"},{"comment":"The lunar daily variation notation Lp is introduced with Eq. (18), but the index p is not defined (it is described in text as p=1,2,3,4). Please add the definition to the text preceding Eq. (18).","section":"Eq. (18) and Section 5.3"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is likely within scope for GJI and the data availability is commendable. The main risk is that the thin-sheet ionosphere approximation, with an unspecified height h, is the linchpin of the separation and is not validated. This is a fixable issue if the authors can add sensitivity studies or at least a clear statement of the assumption's limitations. The lack of error bars is also a concern for a paper that makes many spectral claims."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The headline is simple: this paper delivers something that did not exist before — a decade-long, 3-hourly time series of separated ionospheric, magnetospheric, and induced Gauss coefficients, estimated with no prescribed temporal harmonics and across all local times and magnetic conditions. If the method holds up, it is a useful resource for storm-time external field studies and for mantle induction work. The core idea is sound: ground and satellite data combined with a thin-sheet ionosphere parametrization makes the three-way separation tractable, and the authors implement it carefully with cross-validated truncation, robust loss, and a sensible treatment of observatory biases (including a clear proof of the non-uniqueness in Appendix B). The validation is genuinely good. The qmag10 time series tracks CI and CHAOS during storms, lunar tidal peaks appear only in ionospheric coefficients and not in magnetospheric ones, and the Sq current vortices look right. The C-response comparison — with versus without an ionosphere — is a strong, physically meaningful check: co-modelling the ionosphere makes the transfer functions coherent and monotonic. That is real evidence the separation is working. The soft spots are real but not fatal. The biggest one is Eq. (12), the thin-sheet relation that ties ionospheric coefficients seen from the ground to those seen from satellites. It is load-bearing, yet the sheet height h is never stated, and no sensitivity test or direct validation of the relation is offered. Real ionospheric currents have vertical structure and field-aligned components, especially during storms, and if the relation is violated the inversion can leak ionospheric signal into the induced and magnetospheric coefficients. The authors should state h, test its influence, and justify the approximation against a finite-thickness or more realistic source. That is the main revision I would demand. Secondary issues: no uncertainty estimates for the coefficients, a uniform variance assumption that gets one sentence, no processing code, and several unresolved '??' figure references that make the manuscript feel unfinished. The abstract promises space weather nowcast/forecast benefits, but nothing in the paper demonstrates them; that sentence should be toned down or backed up. Overall, the mathematical structure is clear, the data product is new, and the internal checks are convincing enough that the work deserves refereeing. The thin-sheet concern is serious but addressable, and the rest is polish. I would send this to peer review and ask for a focused revision on h, uncertainties, and completeness. The likely audience — space physicists and EM induction modellers — will cite it if those issues are fixed.","headline":"A genuinely new 10-year, 3-hourly product separating ionospheric, magnetospheric, and induced fields without temporal priors; the central thin-sheet assumption is under-validated but addressable, and the paper deserves a serious referee.","tokens_in":38613,"tokens_out":1948,"would_cite":true,"duration_ms":21441,"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":"This paper claims that combined ground and satellite observations make the ionospheric–magnetospheric–induced field separation tractable, and demonstrates a 10-year, 3-hourly reconstruction.","keywords":["geomagnetic field modelling","ionospheric currents","magnetospheric currents","electromagnetic induction","spherical harmonic analysis","satellite magnetics","lunar daily variations","geomagnetic storms"],"falsifier":"Apply the estimator to synthetic satellite and ground data generated from a three-dimensional ionospheric current model with vertical extent and field-aligned currents; if the thin-sheet assumption is load-bearing, the recovered magnetospheric coefficients will show ionospheric contamination, for example lunar tidal lines appearing in $q^\\mathrm{mag}$.","tokens_in":37505,"feed_emoji":"🛰️","tokens_out":9285,"duration_ms":77939,"temperature":0.7,"pith_summary":"This paper tries to establish that the fast-varying magnetic fields of ionospheric, magnetospheric, and internally induced origin can be separated on a global scale when ground observatory data are combined with low-Earth-orbit satellite data. The key move is to treat the ionosphere as a thin current sheet, which gives a one-to-one relation between the fields it produces above and below the sheet, so the three sources map onto distinct spherical-harmonic coefficients. The authors estimate those coefficients independently in three-hour bins over 2014–2023, producing a continuous record that needs no assumed daily or seasonal rhythms. A reader should care because the separated record isolates storm-time ionospheric dynamics, links magnetospheric periodicities to solar rotation, and yields cleaner electromagnetic transfer functions for probing Earth's interior.","feed_headline":"Ground plus satellite data split 10 years of external magnetic fields","feed_subtitle":"Combining observatory and low-orbit satellite data separates all three sources without assuming periodic behavior.","key_machinery":"The central object is the thin-sheet ionospheric model: a spherical current sheet of radius $a+h$ whose radial magnetic field is continuous. Equation (12) converts the ionospheric external coefficients into the internal coefficients seen from orbit, and substituting that relation into the two potential representations (13)–(14) reduces the unknowns to three independent sets of Gauss coefficients. This linear reparametrization is what makes the combined ground-satellite inversion tractable.","core_discovery":"The paper claims that the internal/external ambiguity is broken by placing satellite observations between the ionosphere and the magnetosphere and modelling the ionosphere as a thin spherical sheet at radius $a+h$. The radial-field continuity condition $B_r|_{r\\to(a+h)^+}=B_r|_{r\\to(a+h)^-}$ gives a one-to-one link between the ionospheric external coefficients seen on the ground and the internal coefficients seen in orbit, so the potential can be rewritten with source-specific Gauss coefficients $(g^\\mathrm{int}, q^\\mathrm{ion}, q^\\mathrm{mag})$ (Eqs. 13–14). Estimating these coefficients per three-hour bin over 2014–2023 yields a continuous, decade-long separation of the three sources.","pith_inferences":["The same geometric argument could be applied to other current layers, such as field-aligned currents, by exploiting observing shells at additional altitudes, provided each layer is thin enough for a sheet approximation.","Because the method needs no prior temporal harmonics, applying it to older satellite missions could extend the source-separated record backwards in time.","The three-hour bin width puts a floor on resolvable dynamics; substorm-scale events shorter than roughly three hours will be smoothed, so a denser satellite constellation would be the natural next test of how much faster the separation can go.","The co-estimated induced coefficients could be fed directly into 3-D mantle conductivity inversions, potentially replacing the quiet-time ionospheric corrections those inversions currently rely on."],"forward_implications":["Because the parametrization imposes no time harmonics, the model can use day-side and storm-time data that most prior external-field models discard.","The 10-year coefficient series separates periodicities by source: lunar daily tidal lines appear in the ionospheric coefficients but not in the magnetospheric ones, a direct check on the separation.","Co-estimating the induced field yields C-responses with higher squared coherence and physically monotonic behavior at periods from 8 hours to 10 days, improving estimates used for mantle conductivity.","The three-hour time-binned construction can be updated as new low-latency data arrive, making it suitable for space-weather nowcasting.","Storm-time reconstructions show the ionospheric equivalent current losing its quiet Sq vortex structure and developing transient high-latitude vortices, indicating the model captures non-periodic storm dynamics."],"supporting_citations":[{"why":"Supplies the thin-sheet ionospheric relation used in Eq. (12) and the comprehensive modelling approach this parametrization builds on.","marker":"Sabaka et al. 2002"},{"why":"Establishes that observatory bias and a piece-wise constant field cannot be uniquely estimated from ground data alone, the non-uniqueness this paper proves and then avoids.","marker":"Langel et al. 1982"},{"why":"Provides the Swarm vector magnetic field measurements that anchor the satellite dataset.","marker":"Olsen et al. 2013"},{"why":"Supplies the CryoSat-2 platform magnetometer data added for better local-time coverage.","marker":"Olsen et al. 2020"},{"why":"Supplies the GRACE-FO magnetometer data that extend the satellite dataset from 2018 onward.","marker":"Stolle et al. 2021"},{"why":"Provides the core and lithospheric field models subtracted from the data and the CI external-field model used for comparison.","marker":"Sabaka et al. 2020"},{"why":"Provides the CHAOS-7 external field time series used for comparison during quiet and storm periods.","marker":"Finlay et al. 2020"},{"why":"Supplies the robust regression procedure used to estimate C-responses from the coefficient time series.","marker":"Olsen 1998"},{"why":"Defines the Chapman phase law used to identify lunar daily variations in the ionospheric coefficients.","marker":"Malin & Chapman 1970"},{"why":"Provides the Sq spherical harmonic analysis framework used to interpret the ionospheric daily variations.","marker":"Schmucker 1999"}],"fun_headline_variants":["10-year split of Earth's external magnetic fields from ground and orbit","Ground + satellite data untangle ionosphere and magnetosphere for a decade","New method separates magnetic field sources over 10 years of storms","Decade of external field separation using observatories and satellites","Ionosphere-magnetosphere split achieved without assuming periodic behavior"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the ionosphere behaves as an infinitely thin current shell; if the real currents spread vertically or flow along magnetic field lines, the clean separation between the three sources fails.","fun_headline_variants_meta":{"raw":{"variants":["10-year split of Earth's external magnetic fields from ground and orbit","Ground + satellite data untangle ionosphere and magnetosphere for a decade","New method separates magnetic field sources over 10 years of storms","Decade of external field separation using observatories and satellites","Ionosphere-magnetosphere split achieved without assuming periodic behavior"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000203,"raw_usage":{"total_tokens":1406,"prompt_tokens":984,"completion_tokens":422,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":600,"completion_tokens_details":{"reasoning_tokens":334}},"tokens_in":600,"tokens_out":422,"duration_ms":4234,"temperature":1.0,"reasoning_tokens":334,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T15:48:00.587349+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Apply the estimator to synthetic satellite and ground data generated from a three-dimensional ionospheric current model with vertical extent and field-aligned currents; if the thin-sheet assumption is load-bearing, the recovered magnetospheric coefficients will show ionospheric contamination, for example lunar tidal lines appearing in $q^\\mathrm{mag}$.","supporting_citations":[],"review_version":1}