Acquire with a manifest
Record source URLs, versions, retrieval times, checksums, and expected schemas for every environmental and biological input.
← All researchMF-PRISM-REEF-2026-03 / interactive guide
THE IDEA, MADE VISIBLE
Missing is not zero.
Follow an environmental record through a strict data pipeline—and see where an unsupported answer must stop.
Technical notePublic-Data Validation Without New Fieldwork
Acquire a dated record
A reproducible pipeline keeps source version, spatial support, units and observation time with the values. This calendar is synthetic, not an acquired public dataset.
- Required daily records
- 84
- Available records
- 76
- Locally reconstructed DHW
- Unavailable
8 days are missing. The strict pipeline refuses to reconstruct DHW from incomplete support.
DHW = (1/7) Σ₈₄ Hᵢ · 1{Hᵢ ≥ 1°C}, only with complete daily support
TRY THIS
Switch between strict and zero-filled modes. The second invents a result; it does not recover the missing observations.
What this experiment represents. Synthetic completeness check. No public bleaching records are acquired and no prediction is fitted. The zero-filled result is shown only to demonstrate the bias caused by an invalid policy.
Reef & Climate · Technical Note
Public-Data Validation Without New Fieldwork
A version-aware acquisition and exposure-audit pipeline for PRISM ReefWatch
Technical noteMF-PRISM-REEF-2026-03 / v0.2
Translate observation-aligned evaluation into a version-aware public-data acquisition and exposure-audit prototype. The offline tests preserve missingness, provenance and chronology rather than silently filling gaps.
- Identifier
- MF-PRISM-REEF-2026-03
- Series
- Reef & Climate
- Edition
- Version 0.2
- Length
- 8 pages
- Reserved DOI
- 10.5281/zenodo.22728281 (record reserved)
01
Plain-language guide
3 minute explanation
What this paper is really saying.
The note turns the validation design into a public-data pipeline that can run without fieldwork while refusing to erase missingness, source versions, or chronology.
public datacoral bleachingdata provenance
A computer-only validation program is possible, but only if public records are treated as evidence with provenance—not as interchangeable spreadsheet rows.
The prototype records what was requested, which version arrived, how temperatures were assigned to reef observations, where coverage is missing, and whether acquisition was blocked. Missing outcome is kept distinct from measured zero.
02
How it works
The argument, without the notation.
The paper’s technical details matter, but the basic route can be understood in three moves.
Align exposure carefully
Map temperature histories to biological observations with explicit units, masks, spatial rules, and complete lookback support.
Audit rather than impute silently
Preserve missing values, blocked downloads, and test-double boundaries in machine-readable validation reports.
03
What it adds
The useful takeaways.
The proposed validation workflow requires no new fieldwork.
Versioning and provenance are first-class parts of the result.
The offline-tested prototype distinguishes missing observations from observed zero bleaching.
04
Scope & limits
What this does—and does not—establish.
Current status
Technical note and offline-tested public-data acquisition prototype. No real bleaching records were acquired or fitted, and no new fieldwork is required.
This technical note documents an acquisition and validation method. It is not a fitted biological model and does not claim new field observations.
No real bleaching records were acquired or fitted in this edition.
Public endpoints, schemas, and source versions may change and must be revalidated.
The note establishes data plumbing and auditability, not a biological forecasting result.
05
Why it matters
Why anyone should care.
The design makes a serious validation program possible from a computer while treating missing data, source versions, spatial assignment, and chronology as first-class evidence.
06
Key terms
The vocabulary, decoded.
These are the terms needed to understand the claim. The full paper uses them more precisely.
Provenance
The documented origin, version, and processing history of a data item.
Missingness
The state in which a value was not observed; it is not equivalent to zero.
Spatial assignment
The rule mapping an environmental measurement to a biological observation location.
Test double
A controlled substitute used to test software behavior without relying on a live external service.
07
Independent review
Where scrutiny should concentrate.
The strongest review is not a general reaction. It tests the steps most capable of changing the conclusion.
- 01Missing outcome versus measured zero; versioned biological records.
- 02Temperature support, spatial assignment, masks and units.
- 03Blocked acquisition and test-double boundaries.
Publication record
Go from explanation to evidence.
This page is a reading guide, not a substitute for the manuscript. The public record links the explanation to the paper, source package, review materials, and persistent identifier.
- Document
- Technical Note
- Review status
- Open for independent review
- Published
- 2026-09-12
- DOI status
- Reserved · 10.5281/zenodo.22728281
- Canonical record
- MF-PRISM-REEF-2026-03
Metriq PRISM Laboratory, Public-Data Validation Without New Fieldwork, Metriq PRISM Laboratory Technical Note MF-PRISM-REEF-2026-03, Version 0.2, 2026. Corresponding contributor: Daniel H. Jeffery, ORCID 0009-0001-1200-6042. DOI: 10.5281/zenodo.22728281.