{"id":"a11fa4a8-e85a-4582-aa70-732fa8d950d5","arxiv_id":"2504.19534","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"A controlled synthetic experiment shows that incomplete solar magnetic field information causes solar wind speed forecast errors of 59-121 km/s within operational models, and that a latitudinal cloud of predictions reduces those errors by 20-77%.","lead":"This paper builds artificial solar magnetic maps with built-in observational blind spots, feeds them into three operational solar wind forecast models, and measures how much the predicted solar wind speed at Earth changes, finding RMSEs of 59-121 km/s. It matters because it quantifies how much imperfect solar observations, not model physics, may drive forecast errors, and tests a simple way to shrink those errors.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 59–121 km/s RMSE and 20–77% mitigation claims rest on one unvalidated AFT/SARG realization and uncalibrated perturbation amplitudes, so they are not yet shown to describe real observational uncertainty.","rationale":"The reader's weakest_assumption is the same as the load-bearing concern I identify: the artificial AFT ground truth and the ad hoc perturbation set. The central claim requires that the synthetic Sun and perturbation ensemble are representative of real observational incompleteness; that condition is the least secure part of the paper. I considered the cross-model MAS comparison (Section 4.3) and the interval-based RMSD mitigation (Section 4.2) as alternative concerns. The MAS comparison is honestly framed as model spread in the text, and the RMSD reduction is largely a consequence of widening the interval, but either issue would mainly affect interpretation rather than the core in-model RMSEs. The synthetic-truth issue, by contrast, directly sets the scale of every quantitative conclusion. Because the authors state the assumption and the design is internally consistent, this is a conditional-acceptance concern, not a rejection; the numbers should not be used to guide operations until a multi-realization or observationally calibrated test is done. A single new experiment comparing RMSE distributions across AFT realizations and perturbation amplitudes would settle whether the quoted values are stable. For these reasons I keep the reader's conditional verdict unchanged.","tokens_in":16270,"tokens_out":9387,"duration_ms":105757,"concrete_test":"Re-run the full 16-map/3-model suite on 10 independent AFT/SARG realizations (varying random seeds and active-region emergence times) and, in the same study, re-scale all perturbation amplitudes by factors 0.5 and 2.0. If the resulting distribution of HUXt/WSA/EUHFORIA RMSEs shifts by more than ~20 km/s relative to the quoted 59/87/121 km/s values, or if the mitigation percentages change materially, the headline numbers are realization- and perturbation-specific rather than a general measure of incomplete magnetic field information.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 2 explicitly states that the AFT/SARG map sequence is 'assumed to represent the actual state of the Sun, i.e., the ground truth', but no comparison with observed magnetograms or flux-transport statistics is provided. AFT is run without data assimilation, and the synthetic active-region catalog, GRaFg random flux (mean 0, sigma 1, 1000 points/hour, Section A.1), and polar-field evolution are all model outputs; nothing ties this single stochastic realization to the real Sun. Every headline number (59, 87, 121 km/s within-model RMSE; 20–77% mitigation) is the product of model sensitivity times the amplitude of the map perturbations, and those amplitudes are not calibrated to measured map-to-map discrepancies. The Gaussian kernels (0.5–3.0 degrees), ±30% sin6 polar scaling, uniform pole filling above 68.5 degrees, and resolution changes are plausible but arbitrary; a different but equally plausible perturbation set would rescale the RMSEs. The single-rotation, single-realization design gives no estimate of realization-to-realization scatter. The paper is transparent about the assumption, but transparency does not establish that the quantitative conclusions transfer to real incomplete magnetic field information.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper constructs a synthetic 'ground truth' solar magnetic field from one 27-day Advective Flux Transport (AFT) model run without data assimilation, generates a set of modified maps designed to mimic far-side aging, polar-field uncertainty, resolution changes, and smoothing, and then feeds these maps into WSA, HUXt, EUHFORIA, and MAS. Comparing the resulting solar wind speed profiles at Earth's nominal latitude, it reports within-model RMSEs of 59, 87, and 121 km/s relative to the ground-truth map run, larger spreads of 77–172 km/s when comparing to MAS runs, and RMSD reductions of 20–77% when predictions are drawn from latitude intervals around Earth. The paper concludes that incomplete magnetic field information contributes roughly 100 km/s of uncertainty in background solar wind forecasts and that multi-latitude prediction intervals can mitigate part of this uncertainty.","tokens_in":16508,"tokens_out":6687,"duration_ms":70470,"significance":"The study addresses an important operational problem with a clean experimental idea: controlled perturbations of input magnetograms can isolate the contribution of incomplete magnetic field information from other forecast error sources, and running three standard operational models under identical inputs is a useful design. The transparent statement of the synthetic 'ground truth' assumption and the comparison against two MAS heating configurations are strengths. If the quantitative values were validated or explicitly treated as a proof-of-concept sensitivity envelope, the approach would be a valuable template for input-uncertainty quantification in solar wind forecasting. As it stands, the specific numbers are conditional on an unvalidated synthetic Sun and on uncalibrated perturbation amplitudes, and the headline mitigation result is partly a consequence of the interval-based metric rather than of improved forecast information.","major_comments":[{"comment":"The paper states that 'we assume that this set of maps represents the actual state of the Sun, i.e., the ground truth' for an AFT run without data assimilation, using SARG active regions and GRaFg random flux (Section A.1). No comparison of this synthetic sequence to observed magnetograms, synoptic maps, or observed flux-transport statistics is provided. Consequently, the RMSE values reported in Sections 4.2 and 4.3 (59–121 km/s within models and 77–172 km/s against MAS) are properties of one stochastic realization of a model Sun, not of real incomplete magnetic field information. The authors should either validate the synthetic maps statistically against observations or explicitly reframe the absolute magnitudes as proof-of-concept sensitivity estimates rather than as observational uncertainties.","section":"Section 2"},{"comment":"The perturbation amplitudes are chosen by hand and are not calibrated to measurements: Gaussian kernel widths of 0.5–3.0 degrees, a ±30% sin6 polar-field scaling, uniform polar filling above 68.5 degrees, and the GRaFg parameters are all plausible but arbitrary. Because the reported RMSEs scale directly with the perturbation amplitude, the headline numbers (59, 87, and 121 km/s; 77–172 km/s against MAS) are not tied to observed map-to-map discrepancies. The manuscript should either calibrate the perturbations to measured uncertainties (for example, differences between contemporaneous maps from different instruments or between synoptic and data-assimilated maps) or report how the results depend on perturbation amplitude, so that the quantitative conclusions are not artifacts of the chosen perturbation magnitudes.","section":"Section 2 and Section 3"},{"comment":"The 'cloud of points' mitigation result is strongly influenced by the definition of the RMSD metric: Eq. (12) assigns zero deviation whenever the ground-truth value falls inside the lower/upper bounds, so widening the latitude interval mechanically reduces RMSD. The reported 20–40%, 38–57%, and 61–77% reductions are therefore partly guaranteed by the metric rather than by genuine forecast skill. Please separate the geometric effect of interval widening from actual predictive improvement, for example by reporting interval coverage probabilities or by testing the latitude-interval predictions against observed solar wind time series rather than against the synthetic ground-truth profile.","section":"Section 4.2 and Table 2, Eq. (12)"},{"comment":"All conclusions rest on a single 27-day AFT realization and a single set of map modifications, so there is no estimate of realization-to-realization or rotation-to-rotation scatter. This matters for the robustness of the qualitative differences between models, such as the reported anti-correlation of EUHFORIA with the ground-truth run in Section 4.2. A multi-rotation or multi-realization analysis, or at least an explicit sensitivity test over independent synthetic realizations, is needed before the values can be treated as general uncertainty estimates for solar wind forecasting.","section":"Sections 2, 4.2, and 4.3"}],"minor_comments":[{"comment":"The map counting is inconsistent: the text says 16 maps including the ground truth map and 15 variations, but Figures 2–3 appear to show only 14 modified maps in addition to the ground truth if the MAS-processed map in Figure 2b is not an operational input. Please clarify exactly which maps were used as input for WSA, HUXt, EUHFORIA, and MAS.","section":"Section 2 and Section 3"},{"comment":"The text gives the HUXt median-vs-ground-truth RMSE as 58.6 km/s, whereas Table 3 lists RMSE = 59.2 km/s and sigma_RMSE = 58.6 km/s; please harmonize the quoted value.","section":"Section 4.2 and Table 3"},{"comment":"The '20–77%' mitigation range is presented as a single headline number, but it corresponds to three different latitude intervals (±4°, ±8°, and ±12°); the abstract and discussion should make this dependence explicit.","section":"Abstract and Section 4.2"},{"comment":"The concluding sentence that increased operational coverage 'could reduce the uncertainty in the background solar wind by approximately 100 km/s near 1 au' is not directly supported by the experimental design, which perturbs existing maps rather than simulating the effect of added far-side or polar observations. Please soften or reframe this claim.","section":"Section 5"},{"comment":"The panel labels and color-bar annotations are very small and difficult to read; larger fonts or separate labeled panels would improve reproducibility and clarity.","section":"Figures 2 and 3"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is within the journal's scope and the experimental approach is useful. The main risk is that the quantitative uncertainty values will be cited as observational error bars even though they are conditional on a synthetic, unvalidated ground truth and on hand-chosen perturbation amplitudes. The revision should either add validation against observations or explicitly demote the numbers to a proof-of-concept sensitivity study. I see no concerns about novelty or attribution beyond the need to tighten the map-count and statistics consistency noted in the minor comments."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First off: this is a useful, honest paper. It isolates the contribution of incomplete magnetic-field information to solar wind forecast error in a controlled way that I haven't seen done across three operational models before. The quantitative headline—59, 87, 121 km/s RMSE for HUXt, WSA, EUHFORIA against a synthetic ground truth—lands in the same range as Poduval et al.'s 85-110 km/s, which tells me the experiment is hitting something real. The paper also does something sensible with a 'cloud of latitudes' around Earth and shows it can capture the reference run more often.\n\nThe main soft spot is the ground truth. The authors assume AFT's single synthetic rotation is the actual Sun. They say so explicitly in Section 2, and no attempt is made to show that this realization statistically resembles observed magnetograms or observed flux-transport behavior. AFT is configured with SARG active regions and GRaFg random flux, and the polar-field evolution is model output. Nothing ties the amplitude of the perturbations to measured map-to-map discrepancies. The Gaussian kernels, the ±30% sin6 polar scaling, the pole-fill threshold—these are reasonable starting points, but they're hand-chosen. Different choices would rescale the RMSEs, and without calibration we only learn that map errors matter, not how much they matter for a given observing system.\n\nThe mitigation result is also weaker than it first appears. Widening the latitude interval mechanically increases the chance that the truth falls inside the interval, so part of the 20-77% reduction is built into the RMSD definition. The fact that the truth rarely falls in the modified-map range is a real and interesting finding, but the 'cloud of points' is tested with knowledge of the truth, which is fine for a proof of concept and not a recipe for operations.\n\nI'd also like to see the maps and code released, and more than one synthetic rotation—the realization-to-realization scatter could be significant.\n\nNone of this kills the paper. The central argument—that incomplete magnetic data contributes tens of km/s of forecast uncertainty and that ensemble ranges need to account for it—holds up. The numbers are plausible and line up with earlier single-model estimates. The paper deserves a serious referee; I'd send it to review and ask for a revised version that either validates the synthetic ground truth against observations or clearly reframes the results as model-sensitivity analysis rather than a measurement of real-world uncertainty.","headline":"A carefully controlled multi-model sensitivity experiment that gives a plausible range for map-induced forecast error, but the absolute numbers rest on an unvalidated synthetic Sun and uncalibrated perturbations.","tokens_in":17070,"tokens_out":2993,"would_cite":true,"duration_ms":30489,"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":"Missing solar magnetic data alone causes 59-121 km/s wind forecast errors","keywords":["solar wind forecasting","heliospheric modeling","solar magnetic field","synoptic charts","surface flux transport","space weather","uncertainty quantification","Carrington rotation"],"falsifier":"Take a real, well-observed Carrington rotation, build the same 15 perturbation maps from actual synoptic magnetograms, run the three forecast models, and check whether the observed in-situ solar wind speed at L1 falls inside the modified-map envelope; if the observed speed frequently falls outside, the synthetic ground truth overstates how much of real forecast error is driven by missing magnetic information.","tokens_in":1954,"feed_emoji":"☀️","tokens_out":3600,"duration_ms":89116,"temperature":0.7,"pith_summary":"This paper asks how much of the error in ambient solar wind forecasts comes specifically from incomplete information about the Sun's magnetic field, rather than from model physics. To isolate that source, the authors generate a synthetic solar rotation with a surface flux transport model, declare one map the ground truth, and create 15 modified maps that mimic far-side aging, polar field errors, resolution limits, and smoothing. They feed all 16 maps into three operational solar wind models—WSA, HUXt, and EUHFORIA—and measure the spread in predicted speed at Earth's latitude. The central result is that incomplete magnetic information alone produces root-mean-square errors of 59, 87, and 121 km/s in the three models, and the spread widens to 77-172 km/s when the models are compared with MHD reference runs. A practical mitigation already visible in the numbers is that predicting a range of speeds from a small latitude band around Earth shrinks the deviation from the ground truth by 20-77%.","feed_headline":"Missing solar magnetic data alone causes 59-121 km/s wind forecast errors","feed_subtitle":"Three operational models differ from their ideal-input run by 59-121 km/s; latitude sampling cuts errors by 20-77%.","key_machinery":"The load-bearing device is a controlled perturbation experiment around a synthetic ground truth. AFT, a surface flux transport model that advects magnetic field by differential rotation, meridional flow, and convective simulations, produces 27 daily full-Sun maps, and the middle map is treated as the truth. From it the authors build one unperturbed and 15 perturbed magnetograms: a synoptic chart assembled from a single viewpoint, a distance-dependent smooth to mimic far-side aging, ±30% pole-field scaling using a $\\sin^6$ function, pole filling above 68.5 degrees, resolution changes by factors of 2, 1/2, and 1/4, and six Gaussian smoothings from 0.5 to 3.0 degrees. Feeding these maps through WSA, HUXt, EUHFORIA, and two MAS heating configurations converts map-level uncertainty into physical units—km/s of solar wind speed at Earth—which is the conversion needed to quantify the problem.","core_discovery":"The paper establishes a controlled error budget: when the only thing changed between runs is the magnetic map used as input, the resulting spread in solar wind speed at Earth's location is comparable in size to the spread seen when different models are run on identical maps. The median modified-map run misses the ground-truth run by 59 km/s for HUXt, 87 km/s for WSA, and 121 km/s for EUHFORIA, and the three models' forecasts differ from the thermodynamic MHD reference by an average of roughly 121-127 km/s. The largest single contributor among the tested perturbations is the synoptic-chart construction, which combines the aging effect with the absence of far-side information. The paper also finds that the ground-truth forecast frequently falls outside the envelope of the perturbed runs, so a naive ensemble of uncertain maps is not guaranteed to bracket the true solution. Consequently, the authors argue, replacing a point forecast with a velocity range derived from a latitude cloud around Earth reduces the chance of missing the ground truth by 20-77%, and better full-Sun magnetic coverage could cut ambient-wind forecast error by roughly 100 km/s near 1 au.","pith_inferences":["Because the perturbation sizes—Gaussian widths of 0.5-3.0 degrees and ±30% polar scaling—are not calibrated to any instrument, the absolute RMSE numbers should be read as order-of-magnitude estimates; a calibration against simultaneous multi-viewpoint magnetograms would either confirm or rescale them.","The latitude-cloud mitigation suggests a cheap operational upgrade: instead of tracing one field line to Earth, issue a probabilistic interval from several closely spaced latitudes, and score it directly against historical in-situ spacecraft data where the true speed is known.","The paper's assumption that the synthetic Sun is the truth means every RMSE measures deviation from an artificial reality; repeating the experiment with a real, well-observed solar rotation degraded to Earth-only coverage would show how well the modified-map envelope brackets actual observed speeds."],"forward_implications":["If the synthetic ground truth is representative, roughly half of the typical 100-120 km/s forecast skill gap at 1 au can be attributed to input magnetic maps rather than to the wind model itself, so better magnetic observations are a direct path to better ambient forecasts.","Operational forecasts that quote a single speed at Earth should be supplemented by a speed interval sampled from a latitude band: the paper's numbers suggest a ±12 degree band cuts the deviation from truth by 61-77%.","Ensembles built only from perturbed magnetograms can miss the true state; the ground-truth run falls outside the modified-map envelope for some longitude ranges, so uncertainty quantification needs model diversity as well as input perturbations.","The synoptic-chart construction is the dominant tested error source, adding at least 10, 40, and 70 km/s for HUXt, WSA, and EUHFORIA, so reducing the aging effect through more frequent full-Sun maps or far-side data should have the largest payoff.","Different models disagree with each other more than any single model disagrees with its own perturbed inputs, so multi-model spread, not within-model ensemble width, is the conservative measure of current forecast uncertainty."],"supporting_citations":[{"why":"Supplies the AFT surface flux transport model that generates the synthetic ground-truth magnetic field maps.","marker":"L. Upton & D. H. Hathaway 2014a"},{"why":"Provides the synthetic active region generator used to populate the artificial solar rotation with realistic bipolar active regions.","marker":"B. K. Jha & L. A. Upton 2024"},{"why":"Defines the WSA model used as one of the three operational solar wind forecast models.","marker":"C. N. Arge & V. J. Pizzo 2000"},{"why":"Defines the HUXt time-dependent solar wind model used for the forecast runs.","marker":"M. J. Owens et al. 2020"},{"why":"Defines the EUHFORIA heliospheric MHD model used for the forecast runs.","marker":"J. Pomoell & S. Poedts 2018"},{"why":"Defines the MAS thermodynamic MHD code whose two heating configurations provide the comparison reference runs.","marker":"Z. Mikić et al. 1999"},{"why":"Documents the synoptic-chart aging effect that motivates the far-side uncertainty construction.","marker":"S. G. Heinemann et al. 2021"},{"why":"Provides the typical forecast-versus-observation RMSE baseline used to judge whether the reported uncertainties are realistic.","marker":"M. A. Reiss et al. 2020"}],"fun_headline_variants":["Solar wind forecast errors from missing magnetic data: 59-121 km/s","Incomplete solar magnetic maps cause 59-121 km/s forecast errors","How much do missing solar magnetic observations hurt wind forecasts?","Magnetic data gaps drive solar wind forecast uncertainty up to 121 km/s","Using latitude cloud cuts solar wind forecast misses by 20-77%"],"cache_read_input_tokens":19200,"weakest_assumption_plain":"The whole uncertainty budget is measured against a synthetic Sun produced by AFT without data assimilation; if that synthetic field does not statistically resemble the real Sun, the RMSE numbers apply only to this artificial realization.","fun_headline_variants_meta":{"raw":{"variants":["Solar wind forecast errors from missing magnetic data: 59-121 km/s","Incomplete solar magnetic maps cause 59-121 km/s forecast errors","How much do missing solar magnetic observations hurt wind forecasts?","Magnetic data gaps drive solar wind forecast uncertainty up to 121 km/s","Using latitude cloud cuts solar wind forecast misses by 20-77%"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000385,"raw_usage":{"total_tokens":2089,"prompt_tokens":1050,"completion_tokens":1039,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":666,"completion_tokens_details":{"reasoning_tokens":944}},"tokens_in":666,"tokens_out":1039,"duration_ms":7715,"temperature":1.0,"reasoning_tokens":944,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T05:50:30.019792+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a real, well-observed Carrington rotation, build the same 15 perturbation maps from actual synoptic magnetograms, run the three forecast models, and check whether the observed in-situ solar wind speed at L1 falls inside the modified-map envelope; if the observed speed frequently falls outside, the synthetic ground truth overstates how much of real forecast error is driven by missing magnetic information.","supporting_citations":[{"cited_title":"G., Temmer , M., Hofmeister , S","cited_arxiv_id":null,"evidence_quote":"Documents the synoptic-chart aging effect that motivates the far-side uncertainty construction."}],"review_version":1}