For the complete documentation index, see llms.txt. This page is also available as Markdown.

AEPs

Appendix D. Example AEPs, Indices & Event Formats

This appendix provides illustrative, non-normative examples of how Assurance & Evidence Packs (AEPs), indices, and event messages are represented on the Nexus Rail.

The goals are to show:

  • How AEPs carry structured evidence (GRIx-based) with EQL/CL, provenance, and uncertainty.

  • How indices are emitted as streaming events and stored in catalogs.

  • How alerts, triggers, and episodes are encoded for RailOps, agents, and external systems.

All examples are simplified for exposition; real deployments would be backed by JSON Schema/OpenAPI, GRIx ontologies, and NXSS validation rules.


D.1 Example AEPs

D.1.1 Planning AEP (EQL3) — Drought & Food Security

This AEP supports seasonal planning (not direct capital triggers) on the East Africa Drought & Food Security Rail. It is issued at EQL3 with mixed observation–model basis and GRF peer review.

api_version: aep.v1
kind: AEP

metadata:
  aep_id: "aep.eaf.drought-food.seasonal.2026M03.v1"
  title: "Horn of Africa Seasonal Drought & Food Security Assessment — March 2026"
  rail_id: "nxsr.eaf.drought-food.v1"
  nrm_profile_id: "nrm.drought.food.household"
  eql_level: "EQL3"
  cl_target: "CL2"
  chronotope_scope_id: "chrono.eaf.drought-food.season.MAM2026"
  int_modules_used:
    - "CLIMATEINT"
    - "GEOINT"
    - "AGROINT"
    - "SOCINT"
  sd z_tags:
    - "sdz.sov.national"
    - "sdz.reg.regional"
  language: "en"
  status: "issued"

data_sources:
  - source_id: "ds.climate.reanalysis.horn.v3"
    provider: "GCRI/ClimateLab"
    sdz_class: "sdz.reg.regional"
    licence: "CC-BY-4.0"
  - source_id: "ds.ndvi.sentinel2.africa.v2"
    provider: "GCRI/EO-Lab"
    sdz_class: "sdz.reg.regional"
    licence: "CC-BY-4.0"
  - source_id: "ds.fsms.household.surveys.2024-2025"
    provider: "National Statistics + WFP"
    sdz_class: "sdz.sov.national"
    licence: "restricted-statistical"
  - source_id: "ds.pastoralist_knowledge.panels.v1"
    provider: "Pastoralist Councils"
    sdz_class: "sdz.sens.indigenous"
    licence: "custom-community-licence"

models_used:
  - model_id: "mdl.drought_spi.v2"
    version: "2.1.3"
    model_card_ref: "https://example.org/modelcards/mdl.drought_spi.v2"
  - model_id: "mdl.livelihood_response.v1"
    version: "1.0.4"
    model_card_ref: "https://example.org/modelcards/mdl.livelihood_response.v1"
  - model_id: "mdl.food_insecurity_projection.v1"
    version: "1.3.0"
    model_card_ref: "https://example.org/modelcards/mdl.food_insecurity_projection.v1"

indices_produced:
  - index_id: "idx.eaf.spi.seasonal.v3"
    index_type: "risk"
    scale: "subnational"
  - index_id: "idx.eaf.food_insecurity_rate.v2"
    index_type: "risk"
    scale: "subnational"
  - index_id: "idx.eaf.humanitarian_need.v1"
    index_type: "systemic_stress"
    scale: "national"

uncertainty_characterisation:
  narrative: >
    Projections of food insecurity carry high uncertainty in conflict-affected
    districts and in areas with sparse NDVI observations due to cloud cover.
    Household survey under-sampling of nomadic populations introduces additional
    structural uncertainty in pastoralist zones.
  structured:
    parameter_uncertainty: "moderate"
    structural_uncertainty: "moderate_high"
    data_gaps:
      - "pastoralist mobility patterns in cross-border clusters"
    sensitivity_to_assumptions:
      - assumption_id: "asmp.humanitarian_response_timeliness"
        qualitative_level: "high"

co_authorship:
  - actor_type: "NCC"
    actor_id: "ncc.univNairobi.risklab"
    role: "lead_modelling_and_analysis"
  - actor_type: "agency"
    actor_id: "ndma.kenya"
    role: "national_drought_expert_review"
  - actor_type: "community"
    actor_id: "council.pastoralist_north_kenya"
    role: "indigenous_knowledge_contributor"
  - actor_type: "rnc"
    actor_id: "rnc.kenya-nexus"
    role: "coordination"

grf_assurance_status:
  status: "peer_reviewed"
  grf_review_id: "grf.rev.aep.eaf.2026-03-15"
  reviewers:
    - "GRF/EastAfrica-Panel-1"
    - "GRF/FoodSecurity-Panel-2"

intended_use:
  decision_classes:
    - "seasonal_planning"
    - "humanitarian_prioritisation"
    - "contingency_budget_discussion"
  exclusions:
    - "no_direct_capital_triggers"
    - "no_parametric_payout_valuation"

issued_at: "2026-03-15T10:00:00Z"
valid_until: "2026-06-30T23:59:59Z"
signatures:
  - signer_id: "did:nexus:ncc.univNairobi"
    signature_type: "ed25519"
    signature_value: "base64..."
  - signer_id: "did:nexus:grf.eaf.panel1"
    signature_type: "ed25519"
    signature_value: "base64..."

D.1.2 Capital-Linked AEP (EQL4) — Drought Facility Trigger

This AEP is prepared for sovereign risk finance; it reaches EQL4, is linked to a specific facility, and is formally anchored via nxproto.aep-anchor.


D.2 Example Index Event Formats

Indices are typically emitted as streaming events (e.g., over Kafka/NATS/MQTT via NXCOMMS) and logged to the Index Catalog for each rail.

We show examples as JSON using a CloudEvents-inspired structure.

D.2.1 Risk Index Event — Seasonal Drought Index


D.2.2 Resilience Index Event — Urban Service Resilience


D.2.3 Systemic Stress Index Event — Climate–Macro Stress


D.3 Example Alert & Trigger Event Formats

D.3.1 Early Warning Alert Event


D.3.2 Capital Trigger Candidate Event


D.3.3 Data Quality / Evidence Degradation Alert


D.4 Episode & Decision Event Formats

D.4.1 Episode Event


D.4.2 Decision Event


D.5 Common Event Envelope & Headers

Across indices, alerts, AEP anchors, episodes, decisions, and agent actions, Nexus recommends a common event envelope, with:

  • specversion, type, source, id, time, subject, datacontenttype (CloudEvents-style).

  • A nexus section carrying rail_id, sd z tag, EQL/CL context, profile/chronotope IDs.

  • A data section carrying payload aligned with GRIx & NXSS.

  • An extensions section for signatures, trace IDs, safety tiers, governance hints.

In practice, these envelopes are hardened by:

  • Machine-readable JSON Schemas and conformance tests (NXSS §3.9).

  • SDZ-aware routing and policy tags via NXCOMMS.

  • Agent safety checks and audit hooks in NXSOS and NXHIVE.

Together with the ontologies in Appendix C and configuration examples in Appendix B, these event formats show how NRM’s semantic, technical, and governance requirements converge into a coherent, machine-verifiable event grammar for the Nexus Rail.

Last updated

Was this helpful?