CSI measures animal-agriculture pressure as a single index.

Headline CSI

This is an illustrative, provisional candidate snapshot assembled from public national and international source material. It has a candidate-track reproducibility certificate, but it is not benchmark-trusted, production-certified or verified-live. The 2026 label is the nowcast’s observed period, not evidence that the full year has been measured. The candidate bundle declares national statistics, Climate TRACE, national and USDA PSD production, ILO-modelled and national labour-force data for the head-count, GHG, animal-product output, feed and employment inputs. Land use uses a provisional geospatial nowcast. The official pasture-area source remains the reconciliation anchor, while current satellite and national land-cover products prevent the land-use signal from going stale. The trade and consumer metrics carry forward from 2023. The fully-measured base period is 2023. These are calculation-bundle declarations, not authenticated production ingestion.

Static candidate release published/as-of 21 August 2026. The 31 December 2026 value stored in the candidate bundle is a period boundary, not a future operational update.

Source status

The free-public-WORLD-2026 bundle is provisional and has 89.09% model confidence. Validation checks: schema: passed; source-hash: passed; nowcast-carry-forward: applied; release-certification: not-run. Its internal 31 December 2026 timestamps are nowcast period-boundary metadata, not evidence of a collection or publication on that future date. The fail-closed public trust summary was last published 22 Aug 2026.

Where the evidence lives

The source registry, publication calendar, freshness and data-quality layer are on Global Optics.

CSI products

Eight sub-index products compose the CSI.

Each sub-index is a first-class benchmark with its own figure, chart, signals and evidence; the CSI is their weighted composition over 101. Open a product for its detail, or follow it through to the source layer in Optics. Every value is derived from declared candidate inputs and remains provisional until authenticated production lineage and the benchmark trusted gate pass.

How the CSI is calculated

The CSI is the weighted composition of eight sub-indexes.

The headline value is composed from the eight first-class sub-indexes above, one per methodology metric, with raw priority weights summing to 101. The table below shows each sub-index’s raw weight, its weight as a fraction over 101, its latest value and its weighted contribution, plus the direct-versus-composed parity status that gates composition mode.

CSI movement shows direction, not moral certainty.

Within the model, a higher CSI score represents a reduction in livestock-related pressure across weighted metrics and a lower score represents an increase. That direction is a calculation convention, not moral certainty or proof of current conditions.

Any production CSI update would need an authenticated source trail showing source identity, reporting period, validation state, anomaly state, rights review and update timestamp.

Methodology version and proof status.

The illustrative headline is calculated under the recorded csi-v1.1.0 candidate methodology and candidate proof chain. That bounded calculation evidence is not an authenticated production release. The per-epoch ledger and composition records carry the csi-v1.0.0 base figures, and the display headline applies the csi-v1.1.0 trade and consumer recalibration on top.

Data restatement, 13 June 2026: the animal-headcount metric now ingests poultry (FAOSTAT element 5112), which a collection filter had previously dropped, so it had excluded chickens. The methodology formula is unchanged; this is a data-collection correction. It lowers the headline from 77.71 to 76.16 and the production base from 70.92 to 69.37 at 2023; the pre-restatement chain is archived verbatim under the proof ledger. The recompute is mathematically consistent: direct and composed CSI match at E8 at every epoch and every proof block re-hashes to its recorded value.

The last published record reports a 4-of-4 candidate-track calculation result. It remains provisional because candidate calculation checks are not the benchmark trusted gate, and the current public readiness record is not ready. Named-human source-rights review, derivative-index clearance, authenticated production lineage and release publishability remain pending. Every CSI and sub-index number carries an explicit data-state label; public hashes can be recomputed with the verification kit without upgrading those claims.