# RMO — observational inputs, provenance and the next solar stage Accepted at R104a, 2026-09-09. The six-row R98–R103 synthesis is already complete; this update does not repeat its calculations. R104a adds input-source guidance, a saveable worksheet and the observational priority below. The numerical solver is unchanged. ## Governing question RMO must report what follows from the available constraints and which additional measurement could distinguish the remaining alternatives. It must not require that every parameter be measured before any useful conditional comparison. It must never present an assumed value as an observational confirmation. ## Input origin and method are recorded separately | Origin | Meaning | Required traceability | |---|---|---| | MEASURED | Recovered from the specific event data; may involve an observation-based diagnostic | Instrument/file, patch, time, observable, processing or inversion, uncertainties and assumptions | | SOURCE_DERIVED | Published estimate or model for this particular event | Article, exact table/figure/equation, method, assumptions and dependencies | | ASSUMED / LITERATURE-PRIOR | Unmeasured quantity supplied by a justified literature scenario or physical assumption | Source, environment/height, adopted range and its rationale, role in the conclusion | The second field states how the value was obtained: direct event quantity, diagnostic estimate, geometric reconstruction, model calculation or assumption. Thus a data-based estimate is not silently called a direct measurement. For every parameter preserve: value/range and units; origin; method; exact source locator; physical environment; error meaning; input dependencies and assumptions; and effect on the conclusion once that effect has been tested. An angle measured in the image plane is not automatically the three-dimensional field-to-normal angle. A reconstruction must name its geometry and underlying inputs. An angle set to 90° is an assumption. If the reconstruction depends on an assumed field direction, that dependence remains in the reported result. Do not assign an error to an inferred angle without a basis. Report outputs as model-dependent whenever their derivation depends on assumed inputs, even when some of the contributing quantities were measured. ## Démoulin & Klein: what the supplied book actually provides Chapter: P. Démoulin and K.-L. Klein, Structuring of the Solar Plasma by the Magnetic Field, in Transport and Energy Conversion in the Heliosphere (2000), printed pp.99–135. Chapter DOI supplied with the discussion: https://doi.org/10.1007/3-540-45166-8_3 . The attached PDF was inspected directly; the online DOI lookup did not return usable content in this session. Table 1 is on printed p.108 / PDF page 114. The accompanying explanation is on printed pp.107–109 / PDF pages 113–115. Values are typical averages based on observations and semi-empirical modelling, with line-of-sight limitations. | Environment | T | Electron density n_e | Field magnitude | |---|---:|---:|---:| | Coronal hole | ≤1 MK | 10¹³ m⁻³ = 10⁷ cm⁻³ | 10⁻⁴ T = 1 G | | Quiet corona | 2 MK | 10¹⁴ m⁻³ = 10⁸ cm⁻³ | 10⁻³ T = 10 G | | Active-region corona | 4 MK | 10¹⁶ m⁻³ = 10¹⁰ cm⁻³ | 10⁻² T = 100 G | These rows do not supply statistical errors, a complete interval for each parameter, a magnetic direction, or a front normal. The coronal-hole temperature is an upper guide, not a 1 MK central value with a fabricated symmetric error. To adopt a range, justify it for the relevant height and structure using event-specific evidence or an explicitly declared scenario. Do not span the quiet/active/hole rows as if they were one jointly allowed event distribution. The n_H column is neutral hydrogen; it must not be read as proton density. Prominence parameters are saved as a reference but are not automatic coronal presets because ionization and coupling assumptions differ. Table 2 (printed pp.111–112 / PDF pages 117–118) supplies order-of-magnitude scales and parameter dependences about T=10⁶ K, N=10¹⁵ m⁻³ and B=10⁻² T. Use V_A ∝ B/√ρ, c_s ∝ √(p/ρ), and beta=2 mu0 p/B² with a consistent composition and pressure closure. Under fixed composition and thermal closure, the table's N and T scalings follow. Do not enter B, density, temperature, V_A, c_s and beta as independent exact numbers. A bounded source-consistency check illustrates why. Taking fully ionized pure hydrogen with n_e=n_p=N and T_e=T_i=T gives rho=m_p N, p=2NkT. At the printed normalization, V_A≈6898 km/s, c_s≈166 km/s, beta≈6.94×10⁻⁴. The printed scales are 1000 km/s, 200 km/s and 4×10⁻⁴. The first two printed speeds, if treated as exact at gamma=5/3, imply beta=0.048 through beta=2 c_s²/(gamma V_A²), not the printed beta scale. This is a reason to recompute the derived state consistently, not to claim the book provides precise model inputs. Alternative composition/pressure conventions must be stated; the worksheet performs no automatic conversion to solver states. ## How every observational result must be presented First state the result and its claim strength. Then provide an input table: | Quantity | Value and ± bounds | Origin and method | Source | Assumptions / dependencies | Effect on conclusion | |---|---|---|---|---|---| | Each input used in that calculation | Preserve units, asymmetry and error semantics | Measured, published-derived or assumed; state any reconstruction | Exact locator | Include geometry, composition, shared inputs and model closure | Tested stability, demonstrated alternative, or not yet assessed | The result must list the exact input record used, not merely the current values in an editor. Predictions must remain outputs: for example, a downstream temperature computed from the selected shock model cannot be reused as an independent observation confirming the same model. A metadata table by itself does not demonstrate robustness; the applicable physical test is still required. Use centre ± absolute/relative error or explicit asymmetric limits. Label hard measurement bounds, declared scenario bounds, quoted undefined errors, and statistical 1σ/2σ/3σ intervals distinctly. A typical average has no supplied error width. Record allowable joint combinations and shared calibration or physical inputs. Use a probabilistic prior only if its distribution is actually specified and justified; otherwise call it a scenario bound. Say robust under declared literature-bounded assumptions only when the result has been established over the full allowed joint domain. A finite parameter grid justifies a sampled-domain statement. Losing a sufficient exclusion test does not construct a competing solution. Excluding fast does not uniquely identify slow. The illustrative B=5–15 G, B=3–20 G and 12 G threshold in the discussion are not RMO results or defaults; no solar threshold is invented. ## Four answers after the R98–R103 synthesis 1. **What is sufficiently established?** The saved method checks support the stated model-family diagnoses, one published slow-polar comparison, partial-input fast exclusion, a constructive fast/slow ambiguity, and its conditional reduction by normal-field/angle bounds. They do not establish that any particular observed B, density or angle is accurate. R102's ±20% normal-speed result and approximate 24.4% strict half-width limit, and R103's total-speed ±20% with angle 30°±5°, are model-specific hard-bound results. Their original centres, compression and Mach ranges remain essential. 2. **Main observational bottlenecks.** Same front, patch and time for every diagnostic; 3D normal and its projection; normal shock-frame plasma flow rather than image speed; density/compression and thermal diagnostics with their emission assumptions; magnetic magnitude and direction; and the joint errors introduced by shared geometry, density and calibration. An independent measurement is useful only if it constrains the remaining alternatives. 3. **First end-to-end event recommended: 13 June 2010 (Ma et al.; Kozarev et al.).** This is a readiness choice based on the saved source audits, not a newly measured event. It has the richest existing RMO preparation: EUV kinematics, auxiliary radio compression, thermal diagnostics and saved emission questions. The new task is to obtain independent, matched inputs from the actual data. The old model-derived temperature and inferred field are not independent observations. Geometry and radio/EUV association remain open; the event is a candidate for Level A, not already certified Level A. The 2011 EIS case has useful Doppler information but weak density constraints; the 2009 stereo case constrains geometry but not the full plasma state; the 2017 SUVI example is valuable but its supplied frames still lack verified timing/calibration. These comparisons use existing saved readiness records. 4. **No further general robustness campaign is mandatory before data work.** Reopen a model check only if the selected event introduces an untested geometry/closure, an uncovered branch boundary or a new claim requiring it. Existing numerical pass results are not proofs of all regimes or global dynamical stability. Actual data processing still needs its own calibration, registration and uncertainty checks; those are part of the observational analysis, not a reason to postpone it with unrelated model experiments. ## Evidence levels and one next bounded action - Level A: maximise independent constraints recovered from the actual event. - Level B: combine partial event constraints with stated literature scenarios. - Level C: diagnose the information needed through scenario/sensitivity cases. These describe evidence coverage, not fast/slow labels. Ten or twenty future events are not a claim that they have already been analysed. The next major stage is observational. Its first bounded action is the same-patch pilot for the 13 June 2010 outer EUV dome: define one front patch and interval around the published 05:40 UT radio comparison, inspect image headers/cadence, select a minimal common AIA cutout, and recover its trajectory and local projected normal independently. A candidate interval is 05:37–05:43 UT with pre-event context around 05:35 UT; these are proposed acquisition limits, not new measurements. State exactly what further geometry is needed for a 3D normal. Prepare a bounded acquisition record under existing data caps before downloads; no observation archive has been downloaded in R104a. Do not substitute a broad imaging speed range for its uncertainty at 05:40 UT. Preserve the radio lane assignment as a model-dependent constraint until association is supported. This pilot is complete when its selected pixels, frame times, front definition, trajectory, projected normal, uncertainty sources and remaining 3D ambiguity are reproducible. It supports the article's observation-to-input chain. It does not by itself classify the shock. E04's earlier thermal-only proposal is now deferred rather than automatically launched. SUVI remains a later optional case. ## Implementation at R104a The QuickLook helper provides per-input origin, method, environment, error format/meaning, source and dependency notes. It offers only the three coronal Table 1 guides, explicitly as assumptions, and protects event-labelled values from table replacement. Blank values stay unknown. It previews a provenance table and saves/restores a JSON worksheet with physical units. No family is assigned; it is not silently inserted into the normalized full-state solver. The following observational stage must carry that provenance into its actual result record. Existing exact model outputs remain labelled as model results. The accepted R104 layout and viewer remain, including frame 23 and GOES/SUVI credit. A copy of the confirmed R104 HTML is retained in checkpoints/R104. A full scientific checkpoint does not promise every historic HTML was separately retained. No local user files were deleted or moved. R73 is not newly requested. Source and evidence files: Demoulin_Klein_2000_registry.json, table2_consistency.json; the supplied immutable book copy; the saved R104 synthesis; and results/solar_diagnosis_audit/RMO_solar_diagnostic_audit.md with later R84–R94 results overriding obsolete implementation limitations.