{"id":"40928bd6-0b2d-4546-be6b-5cc907a535bb","arxiv_id":"2606.27620","paper_version":1,"verdict":"UNVERDICTED","confidence":"LOW","novelty_score":5.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":1,"one_line_summary":"Recurrent reduced-order model reconstructs meridional and equatorial solar-wind velocity and density fields from sparse probe time histories using WSA-ENLIL simulation data.","lead":"The paper introduces a recurrent reduced-order learning framework to reconstruct 2D solar-wind plasma fields from sparse temporal signals of a few virtual probes in simulations. A smart generalist might read it to see how machine learning can help extract spatial information from limited spacecraft measurements in space plasma research.","discovery_kind":"new_method","skeptic_critique":{"model":"grok-4.3","headline":"Framework trained and tested only on WSA-ENLIL simulations; no real spacecraft data validation reported.","rationale":"Reader's weakest assumption correctly flags the low-rank premise inside the simulation. The additional load-bearing issue is the missing transfer step from simulation to real observations, which directly affects applicability to the stated use case of spacecraft data. This does not invalidate the simulation results but keeps the practical claim unverified, consistent with the reader's low-confidence UNVERDICTED status.","tokens_in":1732,"tokens_out":333,"duration_ms":24269,"concrete_test":"Train the reported architecture on WSA-ENLIL, then feed real time series from co-located probes (e.g., ACE + STEREO-A for a selected Carrington rotation); compare reconstructed meridional/equatorial fields against independent tomographic or multi-spacecraft reconstructions and report whether RMS error in velocity or density exceeds the simulation hold-out error by more than 25%.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim states that from sparse temporal probe signals the model recovers dominant radial/latitudinal variations and spiral organization, enabling inference from spacecraft measurements. All quantitative demonstrations, including recovery of coupled fields and sensitivity to modal rank, probe count, and history length, use virtual probes placed inside the WSA-ENLIL domain. Real measurements introduce instrument noise, data gaps, calibration offsets, and statistical differences from the model's steady-state assumptions and inner-boundary conditions. No domain-shift test or observational validation is described, so it is unclear whether the low-rank recurrent mapping remains accurate when the input statistics deviate from the training simulation.","agreement_with_reader":"partial"},"referee_report":{"model":"grok-4.3","summary":"The paper presents a recurrent reduced-order learning framework trained on WSA-ENLIL solar-wind simulations to reconstruct 2D meridional and equatorial plasma fields (radial velocity and density) from sparse temporal signals of a small number of virtual probes. It claims recovery of dominant radial/latitudinal variations and equatorial spiral organization, reconstruction of dynamically coupled fields, and sensitivity of accuracy to modal rank, probe count, and input history length, positioning the method as a route to infer spatial structures from spacecraft measurements.","tokens_in":1855,"tokens_out":417,"duration_ms":12937,"significance":"If the reconstruction accuracy holds under broader testing, the approach could offer a practical data-driven tool for interpreting limited in-situ measurements in heliophysics by leveraging reduced-order representations and recurrent mapping. The demonstration on coupled fields and parameter sensitivities is a positive step, but the exclusive reliance on simulation data limits immediate applicability to observations.","major_comments":[{"comment":"Abstract and results (sensitivity studies paragraph): The central claim that the model 'recovers the dominant radial and latitudinal variations' and 'spiral-shaped organization' is stated without any reported quantitative metrics (e.g., RMSE, correlation, or normalized error) or error analysis for the reconstructions, despite explicit sensitivity studies on modal rank, probe count, and history length. This absence prevents assessment of whether the low-rank recurrent mapping actually achieves the claimed recovery.","section":"Abstract"},{"comment":"Abstract (final sentence) and discussion of spacecraft application: The assertion that the methodology provides 'a practical route for extracting spatial plasma-state information from spacecraft measurements' is load-bearing for the paper's motivation but rests entirely on virtual probes inside the WSA-ENLIL domain. No domain-shift tests, instrument noise injection, or real spacecraft data validation are described, leaving open whether the learned mapping generalizes when input statistics deviate from the steady-state simulation assumptions.","section":"Abstract"}],"minor_comments":[],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":{"model":"grok-4.3","summary":"We thank the referee for the constructive comments. We address each major point below and indicate where revisions will be made.","responses":[{"response":"The sensitivity studies section quantifies reconstruction performance via error metrics (including RMSE) as functions of modal rank, probe count, and history length, with results shown in the associated figures. However, these quantitative values are not summarized in the abstract. We will revise the abstract to include explicit metrics (e.g., typical RMSE ranges) supporting the recovery claims.","revision_made":"yes","referee_comment":"[Abstract] Abstract and results (sensitivity studies paragraph): The central claim that the model 'recovers the dominant radial and latitudinal variations' and 'spiral-shaped organization' is stated without any reported quantitative metrics (e.g., RMSE, correlation, or normalized error) or error analysis for the reconstructions, despite explicit sensitivity studies on modal rank, probe count, and history length. This absence prevents assessment of whether the low-rank recurrent mapping actually achieves the claimed recovery."},{"response":"We agree that the demonstration uses only virtual probes from the steady-state WSA-ENLIL simulations and does not include domain-shift, noise, or real-data tests. This is an inherent limitation of the present proof-of-concept study. We will revise the abstract's final sentence and the discussion to clarify that the work establishes the method on simulation data as a foundation, while noting the need for future observational validation.","revision_made":"yes","referee_comment":"[Abstract] Abstract (final sentence) and discussion of spacecraft application: The assertion that the methodology provides 'a practical route for extracting spatial plasma-state information from spacecraft measurements' is load-bearing for the paper's motivation but rests entirely on virtual probes inside the WSA-ENLIL domain. No domain-shift tests, instrument noise injection, or real spacecraft data validation are described, leaving open whether the learned mapping generalizes when input statistics deviate from the steady-state simulation assumptions."}],"tokens_in":1410,"tokens_out":428,"duration_ms":35362,"standing_objections":[]},"desk_editor":{"model":"grok-4.3","letter":"The paper introduces a recurrent reduced-order learning framework that takes time histories from a small number of virtual probes and outputs 2D maps of radial velocity and density, recovering the main radial and latitudinal structure in the meridional plane plus the spiral pattern in the equatorial plane.\n\nThe combination of recurrent networks with reduced-order modeling for this specific solar-wind task is new, and the work shows the model can also fill in coupled fields that were not directly measured. The sensitivity checks on modal rank, probe count, and input history length are useful for understanding behavior inside the training data.\n\nEverything is trained and tested on WSA-ENLIL simulation outputs. No real spacecraft measurements appear, so the method has not been checked against instrument noise, data gaps, or differences between the simulation's steady-state assumptions and actual observations. The abstract gives no error numbers or validation scores, which makes it difficult to assess how accurate the reconstructions actually are.\n\nThe low-rank assumption works for the simulation fields, but it is unclear how well it survives when the input statistics change. This limits how far the claims about practical use with spacecraft data can be taken right now.\n\nThe paper is aimed at heliophysics researchers who need ways to infer spatial plasma structure from limited probe time series. A reader working on data-driven methods for space plasmas would see a concrete demonstration, but the simulation-only scope keeps the impact contained.\n\nIt deserves peer review so that referees can ask for quantitative metrics and at least one test against real observations.","headline":"Recurrent reduced-order model reconstructs 2D solar-wind fields from sparse virtual probes inside WSA-ENLIL runs, but stays simulation-only with no quantitative metrics shown.","tokens_in":2349,"tokens_out":385,"would_cite":false,"duration_ms":22240,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.3","headline":"A recurrent reduced-order model reconstructs solar-wind plasma fields from sparse probe time series.","keywords":["solar wind","plasma reconstruction","reduced-order modeling","recurrent learning","spacecraft measurements","heliospheric structures","meridional plane","equatorial solar wind"],"falsifier":"A test in which reconstructed fields from the learned low-rank model show large pointwise or structural deviations from independent high-resolution simulation snapshots or from real spacecraft observations when the same sparse probe inputs are supplied.","tokens_in":2617,"feed_emoji":"☀️","tokens_out":669,"duration_ms":30905,"temperature":0.7,"pith_summary":"The paper develops a recurrent reduced-order learning framework that takes time histories from a small number of virtual probes as input and produces reconstructed two-dimensional fields of radial velocity and plasma density. Trained on WSA-ENLIL simulation data, the approach recovers the main radial and latitudinal patterns in the meridional plane together with the spiral organization in the equatorial plane. It also infers spatial distributions of plasma quantities that are dynamically coupled but not directly measured by the probes. Sensitivity tests examine how accuracy changes with modal rank, number of probes, and length of input history. The method is positioned as a way to obtain needed spatial context from the limited local sampling that spacecraft provide.","feed_headline":"Recurrent model recovers solar-wind fields from sparse probes","feed_subtitle":"Time histories from a few virtual probes yield reconstructed 2D velocity and density maps in meridional and equatorial planes.","key_machinery":"The recurrent reduced-order learning framework, which learns a low-rank spatial representation from simulation data and uses recurrent mapping to convert probe time histories into the coefficients of that representation.","core_discovery":"From sparse temporal probe signals as inputs, the recurrent reduced-order model recovers the dominant radial and latitudinal variations in the meridional plane and the spiral-shaped organization of the equatorial solar wind, while also reconstructing spatial distributions of dynamically coupled plasma fields not directly sensed.","pith_inferences":["The same low-rank recurrent mapping could be retrained or fine-tuned on actual multi-spacecraft observations to test performance on real rather than simulated data.","Success in recovering unsensed coupled fields suggests the approach may allow inference of additional plasma parameters from the same limited probe set.","The reported dependence of accuracy on probe count and history length supplies concrete guidance for choosing sampling strategies in future heliospheric missions."],"forward_implications":["The model reconstructs spatial distributions of dynamically coupled plasma fields not directly sensed by the probes.","Reconstruction accuracy depends on modal rank, number of probes, and input-history length, as quantified in the sensitivity studies.","The framework supplies a practical route for extracting spatial plasma-state information from spacecraft measurements.","It supports studies on the underlying physics of space and solar-wind plasmas by providing the needed spatial context."],"fun_headline_variants":["Recurrent reduced-order model maps solar-wind fields from probes","Sparse probe signals yield 2D meridional and equatorial plasma maps","Reduced-order learning recovers solar-wind structures from trajectories","Probe histories reconstruct spiral equatorial and radial solar wind"],"cache_read_input_tokens":2112,"weakest_assumption_plain":"The solar-wind plasma structures in the target quantities are sufficiently captured by a low-rank reduced-order representation learned from the WSA-ENLIL simulation data, allowing recurrent mapping from sparse time histories to spatial fields.","fun_headline_variants_meta":{"raw":{"variants":["Recurrent reduced-order model maps solar-wind fields from probes","Sparse probe signals yield 2D meridional and equatorial plasma maps","Reduced-order learning recovers solar-wind structures from trajectories","Probe histories reconstruct spiral equatorial and radial solar wind"]},"model":"grok-4.3","cost_usd":0.002467,"raw_usage":{"total_tokens":1418,"prompt_tokens":647,"num_sources_used":0,"completion_tokens":64,"cost_in_usd_ticks":24674500,"prompt_tokens_details":{"text_tokens":647,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":707,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":647,"tokens_out":64,"duration_ms":9591,"temperature":1.0,"reasoning_tokens":707,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-06-29T01:06:45.719659+00:00","model_set":{"reader":"grok-4.3"},"falsifier":"A test in which reconstructed fields from the learned low-rank model show large pointwise or structural deviations from independent high-resolution simulation snapshots or from real spacecraft observations when the same sparse probe inputs are supplied.","supporting_citations":[],"review_version":1}