ECSS Space Environment — Experimental Irradiation Data in Background Assessments (space-systems/ecss/e1012-bg-experimental)
Use when the task is to incorporate experimental irradiation measurements into a background radiation assessment under ECSS-E-ST-10C §10.4.8 -- validating the experimental records, comparing them against the applicable reference environment model, deciding whether they supersede or supplement that model, and computing a combined background value with a propagated uncertainty envelope.
Domain quick reference
- ECSS-E-ST-10C §10.4.8 permits experimental irradiation data (in-orbit measurements, ground-based irradiation campaign results, or dedicated flight-instrument datasets) to replace or supplement a standard environment model (e.g. AP8, AE8, CREME96) in the background dose or flux assessment, provided the data meet quality and provenance requirements.
- An experimental record must specify: the originating source and instrument, the particle type (proton, electron, heavy ion, or gamma), the radiometric quantity (flux, fluence, dose rate, or dose), the measured value with a relative uncertainty in percent, the measurement duration, a quality flag (valid, suspect, or invalid), and a provenance description linking the data to a documented dataset or campaign report.
- Supersedure of the reference model is granted only when all three criteria hold: quality flag is "valid", relative uncertainty does not exceed MAX_SUPERSEDE_UNCERTAINTY_PCT (20 %), and the signed offset from the reference model stays within ±MAX_OFFSET_FOR_SUPERSEDE_PCT (50 %). Data that pass quality but fail an uncertainty or offset criterion supplement the model through an inverse-variance weighted combination rather than replacing it.
- The combined background value is derived by inverse-variance weighting of the experimental and model estimates on their absolute uncertainties; the reference model carries a default assumed uncertainty of DEFAULT_MODEL_UNCERTAINTY_PCT (30 %) when no instrument-specific value is on record.
- Every experimental record must carry a non-empty source identifier, a non-empty provenance description, and a positive measurement duration. A record with incomplete provenance is flagged in the assessment output even when its numerical values are otherwise compliant.
Workflow
- Collect all experimental irradiation records to be incorporated and validate each one: check that all required fields are present, that particle type and radiometric quantity are drawn from the permitted vocabularies, that the measured value is non-negative, that relative uncertainty is in (0, 100] %, that measurement duration is positive, that the quality flag is one of {"valid", "suspect", "invalid"}, and that source and provenance are non-empty strings. Reject any record that fails validation before it enters the assessment.
- Collect reference environment model records for the same particle-type and quantity combinations. Validate each reference record: positive value, non-empty model name, and compatible particle type and quantity vocabulary. Flag any experimental record whose (particle_type, quantity) pair has no matching reference entry -- it will be carried forward with its own value as the combined result and an explicit "no reference model" issue flag.
- For each matched (particle_type, quantity) pair, compute the signed percentage offset of the experimental value from the reference model value: (experimental − reference) / reference × 100. This offset quantifies systematic agreement and is a prerequisite for the supersedure decision.
- Apply the supersedure criteria to each pair: the experimental data supersede the model only when quality_flag is "valid", uncertainty_pct ≤ 20 %, and |offset_pct| ≤ 50 %. Record the boolean outcome and the rationale string for the assessment report.
- Compute the combined background assessment value and uncertainty:
- If supersedure holds, the combined value and uncertainty equal the experimental value and uncertainty directly.
- Otherwise, apply inverse-variance weighting using absolute uncertainties (sigma_i = relative_unc_i × value_i); the combined value is the weighted mean and the combined uncertainty is propagated from the total inverse variance. The reference model uncertainty defaults to 30 % when not otherwise specified.
- For each record, confirm provenance completeness: source, provenance description, and positive measurement duration must all be present. A record that passes numerical validation but has incomplete provenance receives an explicit provenance issue flag in the output; it is not silently treated as compliant.
- Aggregate the results per (particle_type, quantity) pair into the assessment report: experimental and reference values, model name, offset, supersedure flag and rationale, combined value and uncertainty, provenance status, and any open issues. The pair is fully compliant only when the issues list is empty.
Pitfalls
- Using experimental data that carry a "suspect" or "invalid" quality flag directly in the assessment without applying the supersedure gate -- only "valid" data may supersede a reference model; suspect or invalid data must never replace the model estimate unchecked.
- Treating an offset within ±50 % of the reference model as automatically acceptable without also checking uncertainty -- a high-uncertainty dataset with a plausible offset does not supersede the model; it supplements it through the weighted combination.
- Applying the reference model value as zero uncertainty when computing the weighted average -- environment models carry inherent uncertainty (typically 30 % or more); collapsing that to zero would artificially bias the combined result toward the experimental value even when the experimental uncertainty is large.
- Accepting a record as provenance-complete because it passes numerical validation without separately confirming that a non-empty provenance description and source identifier are on record -- incomplete provenance is itself a finding under §10.4.8, not a silent pass.
- Comparing experimental and reference records across different particle types or radiometric quantities -- offset computation requires an identical (particle_type, quantity) pair; mismatched pairs must be rejected before any numerical comparison is attempted.
Behavior contract (gate 3)
The validation, offset computation, supersedure decision, combined assessment, provenance check, and full-run orchestration logic are exercised by the gate 3 contract test: scripts/test_e1012_bg_experimental.py against scripts/e1012_bg_experimental_logic.py (stdlib unittest, offline). Run: python3 scripts/test_e1012_bg_experimental.py
Compliance
- ECSS standards are freely downloadable (ESA); cite the source and paraphrase per standards-map.yaml.
- compliance: STANDARDS-REF, gated: false.
