A conformant system declares the basis on which it operates, including the parts where no basis exists. It does not assert compliance with rules that do not apply to it, and it does not assert that rules exist where they do not.
This part exists because of a specific fact. There is no ITU Radio Regulations Article 21 space-to-Earth power flux-density row for 5.8 GHz. A system beaming power there is not in violation of that row; it is outside it. Those are different things, and a schema with no way to express the difference forces every implementer to pick a row that does not govern them and write it down as though it did.
The design insight: a schema with no way to say “there is no rule here” forces every implementer to lie. So the attestation has a positive half and a negative half, and the negative half is mandatory.
spectrum {
jurisdiction, // ISO 3166-1 alpha-2 (+ subdivision)
authority, // "FCC", "Ofcom", "MIC", ...
instrument: 'licence' | 'experimental_authorisation'
| 'special_temporary_authority' | 'exemption'
| 'notification' | 'under_application' | 'none_claimed',
instrumentRef, issuedOn, expiresOn,
band { centre_GHz, occupied_bandwidth_Hz, emissionDesignator? },
allocationStatus: 'primary' | 'secondary'
| 'ism_not_allocated_for_this_use' | 'none',
ituFilingRef?,
article21 {
applicability: 'exact' | 'analogue' | 'none',
rowRef: string | null,
declaredPfd_dBW_m2, refBandwidthHz, arrivalAngleDeg,
marginDb, basis: 'screening' | 'modelled' | 'measured'
}
}
[R-A-001] article21.applicability SHALL be exact only where a row of Article 21
covers this band and service. Where the nearest tabulated row is used as a yardstick,
it SHALL be analogue. Where no row applies, it SHALL be none.
[R-A-002] article21.rowRef SHALL be null when applicability is none, and
SHALL name the row otherwise. A null here is a legal, expected value.
[R-A-005] Where applicability is none, marginDb SHALL also be null.
A margin is a distance from a limit. Declaring that no limit applies and then
reporting a distance from one is incoherent, and an earlier draft of this document did
exactly that in its own worked example. If you screened against the nearest tabulated
row, the applicability is analogue and the row is named — that is what “used as a
yardstick” means. none is for a system reporting no screening at all.
[R-A-003] refBandwidthHz SHALL be present whenever declaredPfd_dBW_m2 is
present, per [R-C-008]. There is no exception.
[R-A-004] marginDb MAY be negative. A negative margin is a disclosure, not an
error, and SHALL NOT be suppressed or clamped.
exposure {
guideline: 'ICNIRP-2020' | 'IEEE-C95.1-2019' | 'national',
nationalRef?,
limitClass: 'general_public' | 'occupational',
limit_W_m2, averagingTimeS, averagingAreaCm2,
peakToMeanFactor, peakToMeanBasis,
declaredPeak_W_m2, ratioToLimit,
assessmentType: 'screening' | 'modelled' | 'measured',
assessedBy { org, role, licensed },
signed, reportRef?,
localExposureBasis: 'SAR' | 'absorbed_power_density' | 'not_assessed'
}
[R-A-010] limit_W_m2 SHALL be accompanied by averagingTimeS and
averagingAreaCm2, per [R-C-009].
[R-A-011] peakToMeanFactor SHALL be accompanied by peakToMeanBasis stating how
it was obtained — for example, “2.5584, from a 10 dB Gaussian edge taper”.
A beam is not a top hat. Illuminating an aperture uniformly spills a large fraction of the power past its edge, so any real design tapers. An assessment that compares an aperture-average intensity against a limit written for the field a person stands in has compared the wrong two numbers.
[R-A-012] assessmentType: 'screening' SHALL NOT be used to support an L3
conformance claim. L3 requires modelled or measured, signed, with a named assessor.
[R-A-013] localExposureBasis SHALL be declared.
A known gap, disclosed rather than hidden. ICNIRP 2020 switches the local-exposure basis at 6 GHz: below it, specific absorption rate; above it, absorbed power density. A system operating at 5.8 GHz is therefore on the SAR side, and a power-density screening does not address local exposure at all. Declaring
not_assessedis permitted and honest; claiming compliance on a power-density basis alone would not be. SeeOBJECTIONS.md§O-5.
environmentalReview { regime, status, ref, challengesPending }
aviation { coordinatingAuthority, restrictionType, ref,
notamProcedure, externalAbortChannel }
externalAbortAuthorities[] { org, channel, credentialRef }
[R-A-020] Any party listed in externalAbortAuthorities[] SHALL have a working
abort channel per [R-S-050].
unclaimed[] — the negative declaration[R-A-030] unclaimed[] SHALL be present and SHALL enumerate every regulatory basis
the operator is explicitly not claiming.
[R-A-031] operatingPosture SHALL be one of authorised,
authorised_experimental, or
operating_without_specific_authorisation_by_declaration.
For a 5.8 GHz space-to-Earth system as of 2026-08, an honest bundle reads:
"unclaimed": [
"itu-rr-art21-space-to-earth-pfd-limit-for-this-band",
"primary-allocation-for-space-to-earth-wireless-power-transfer",
"itu-r-recommendation-governing-this-application",
"national-type-approval-for-transmitter-at-rated-power"
],
"operatingPosture": "operating_without_specific_authorisation_by_declaration"
This does three things nothing else in the specification does.
unclaimed[] arrays of every declared system and see the
shape of the gap in a single query. That is precisely the evidence an ITU-R work-item
proposal needs, and precisely what nobody currently has.[R-A-032] An empty unclaimed[] SHALL be accompanied by operatingPosture:
'authorised' and a resolvable instrumentRef for every applicable regime. An empty
array is a strong claim and is treated as one.