Publication standards
Editorial and research policy
Contrail Research publishes evidence-led atmospheric education and experimental model interpretation. This policy explains how we separate observation from inference, choose sources, disclose limitations, correct errors, and keep commercial considerations out of scientific conclusions.
Mission and scope
Our goal is to make contrail and upper-air cloud research understandable, testable, and correctable. We publish educational guides, documented methods, data products, evidence audits, and interfaces that help readers compare flight and weather context with independent observations.
We do not present the site as:
- an operational aviation-safety, navigation, or flight-dispatch service;
- a certified meteorological forecast;
- a remote chemical-detection instrument;
- proof of a specific aircraft’s intent, payload, or attribution based on map proximity;
- medical, legal, or regulatory advice.
Use of this publication
ContrailResearch.org publishes reporting, education, analysis, commentary, and opinion. Observations, source material, and model interpretations are labeled by type. Content is for general information and does not replace legal, medical, regulatory, aviation-safety, or other professional advice. Check primary sources before acting on it.
Evidence categories and language
Every substantive finding should make clear which level it reaches:
| Label | Meaning | Minimum expectation |
|---|---|---|
| Observed | A source directly records the stated feature or measurement. | Identified instrument/observer, time, place, unit or image provenance, and relevant quality limits. |
| Supported | Multiple independent lines of evidence favor an explanation. | A documented chain linking observation, timing, geometry, atmospheric plausibility, and alternatives. |
| Consistent with | The evidence does not contradict a hypothesis but does not uniquely establish it. | At least one explicit alternative and the missing evidence needed to decide. |
| Unresolved | Available evidence is insufficient or materially conflicting. | A plain explanation of the data gap rather than forced classification. |
| Not supported | A claim lacks necessary evidence or conflicts with stronger evidence. | Citations to the relevant record and a distinction between “not supported” and “impossible.” |
Model “scores” are described as relative indicators unless a published calibration study demonstrates a probabilistic interpretation. Correlation on a map is never silently upgraded to causation.
Source hierarchy
- Primary observations and instrument records: original satellite/radar products, archived research-flight logs, documented measurements, and original datasets.
- Authoritative assessments: IPCC and National Academies reports that evaluate a body of evidence and state confidence.
- Peer-reviewed research: especially open methods, data, uncertainty, and replication or comparison studies.
- Government and institutional explainers: FAA, EPA, NASA, NOAA, NWS, and comparable scientific agencies.
- Provider documentation: specifications, coverage statements, API fields, timestamps, and known service limits.
- Secondary reporting: used for context or discovery, not as the sole basis for a technical claim when a primary source exists.
We link as close as possible to the supporting page, paper, dataset, or documentation. A source’s authority does not remove the need to report its date, scope, assumptions, or uncertainty. We do not cite search-result snippets as evidence.
Drafting and review process
Define the reader’s question
The draft states a specific question and separates background education from a new project finding.
Build a claim ledger
Quantitative and potentially disputed claims are mapped to sources. Original analysis identifies the input record and inference steps.
Check competing explanations
Reviewers look for timing errors, altitude mismatches, incomplete flight coverage, weather-grid limitations, natural-cloud alternatives, and overprecise language.
Check safety and privacy
Public observations are reviewed for precise home locations, faces, license plates, personal identifiers, and operationally sensitive detail.
Publish provenance
The page names its authoring unit, review date, primary sources, and limitations.
A page can be updated for broken links or readability without changing its conclusion. Substantive changes to data, method, conclusion, or confidence require a new reviewed date and a correction/update note.
Corrections and challenges
Send a correction to info@contrailresearch.org with the page URL, disputed text, supporting source, and the correction you propose. We prioritize errors that change safety, attribution, numerical results, source meaning, or the strength of a conclusion.
| Change | Treatment |
|---|---|
| Typographic or formatting fix | Corrected in place; no notice unless meaning could have changed. |
| Clarification | Text improved without changing the finding; reviewed date may be updated. |
| Material correction | Dated note explains the prior error, corrected statement, and effect on the conclusion. |
| Retraction | Page remains available with a prominent notice when its central conclusion is unreliable. |
| Unresolved challenge | Added to the limitations or update log when credible evidence exists but review is incomplete. |
Advertising, funding, and commercial independence
Advertising supports site operations but does not determine scientific claims, model scores, source selection, corrections, or coverage. Advertisements are visually and technically separate from editorial content. Payment does not purchase a favorable finding or removal of a supported criticism.
- Sponsored material, if introduced, must be labeled before the reader opens it.
- Affiliate or referral relationships must be disclosed adjacent to affected links.
- Advertisers do not receive unpublished SkyPulse observation data or editorial pre-approval.
- Pages should contain substantial original value independent of any advertisement.
- Potential conflicts involving a source, vendor, funder, or subject are disclosed on the affected page.
User observations and community evidence
SkyPulse reports are valuable because they add time- and place-bound ground context. They remain user observations unless independently reviewed. We preserve the difference between what a participant selected, what automated image analysis suggested, and what an editor concluded.
- Location is published only at the precision necessary for the research purpose.
- Original timestamps and metadata are retained only under the applicable privacy notice and consent.
- Uploaded images are not treated as chemical measurements.
- Duplicate, manipulated, unsafe, or unverifiable submissions may be excluded from analysis.
- Aggregate findings report selection bias, geographic gaps, and the number of observations excluded.
AI, automation, and generated content
Software may assist with outlining, language editing, transcription, code, data transformation, or candidate classification. Automated output is not a source and does not validate itself. Public scientific pages require human review for factual claims, citations, uncertainty, unsupported implications, and privacy.
We will not:
- publish large numbers of lightly varied pages to capture search traffic;
- invent observations, quotations, author credentials, case-study dates, or citations;
- describe generated or simulated imagery as an observation;
- hide a material automated role in a research conclusion;
- use an AI confidence label as a calibrated probability without validation.
Safety, dignity, and uncertainty
Research should not encourage unsafe photography, trespass, interference with aircraft operations, or the identification and harassment of crew or passengers. Callsigns and public flight data are analyzed at the level necessary to test an atmospheric hypothesis. We avoid sensational framing and do not imply wrongdoing from an ordinary contrail, unusual route, missing flight record, or unresolved observation.