Master Art of Translation | The Seismic Monitoring, Waveform Archives & Earthquake Catalogue Translation System — How FDSN, StationXML, miniSEED, QuakeML and Revisions Keep Their Meaning

How do you translate seismic monitoring records, earthquake catalogue entries, waveform files, station metadata and QuakeML event reports without silently changing the scientific evidence? Accurate seismic data translation means preserving the difference between the earthquake and the recording, the distinction between magnitude and shaking intensity, the exact identities of channels and stations, the time and depth reference of every observation, and the revision status of the event solution. A fluent translation can still be technically wrong if it turns an automatic hypocentre into a final determination or changes raw instrument counts into measured ground acceleration.

Readers searching for seismology translation, earthquake catalogue translation, FDSN StationXML terminology, miniSEED 3 waveform translation, QuakeML event metadata, seismic station identifiers, earthquake magnitude types, phase pick interpretation and seismic data quality assurance need more than a bilingual glossary. They need a method for connecting words to records, epochs, units, network identifiers, waveforms, origins, evaluations, provenance and downstream scientific decisions. This chapter explains how that method works, why apparently small word choices break reproducibility, and how translators and reviewers can detect the damage before release.

The repeatable answer is to establish the record layer first, lock every identifier and measurement to its source, translate only the human-readable interpretation, then run a second examination against both the instrument metadata and the scientific claim. EDKSG-TRANS-MASTER-WORLD-680 is the next reserved-tens world-facing child of Master Art of Translation | The Complete System for Moving Meaning Between Languages. It follows search intent, hidden problem, mechanism, repeatable method, failure modes, worked diagnostics, transfer and next routes. The audience includes researchers, technical editors, humanitarian communicators, data archivists, translation teams, teachers and curious readers anywhere.

Scope and caution. All named stations, events, times and sample records in the exercises are fictional illustrations, not reports of actual earthquakes. This is a guide to interpreting and translating records, not a live hazard warning or a substitute for an authoritative seismic network. An event may be revised repeatedly; never turn an illustration or preliminary computation into a safety instruction. Technical references were checked for this chapter on 31 October 2026, including the FDSN specifications, International Seismological Centre resources, USGS earthquake catalogue documentation and ISO 5060:2024.

Contents and navigation

search intent · hidden problem · mechanisms · method · failure modes · diagnostics · field workshops · practice · glossary · faqs · sources

Search intent: what the reader actually needs to get right

Someone asking how to translate an earthquake catalogue is not simply asking how to translate the word earthquake. One reader may need to understand a public event summary; another may be aligning a multiagency event database; a third may be trying to reproduce a processing chain from raw waveform records. They have different permissible simplifications. A plain-language note can describe why an earthquake was reported as preliminary, but a data handover cannot substitute a casual word for the exact evaluation-status value stored in a machine-readable file. Translators must identify which layer carries the authority and which layer is allowed to be explanatory.

The high-intent questions are practical: What is the difference between origin time and update time? Are station codes proper names? Do digital counts equal ground velocity? Can mb and Mw be translated as the same magnitude? What does a missing waveform segment mean? Which StationXML epoch belongs to the requested trace? How do source identifiers survive migration from legacy channel names to FDSN Source Identifiers? Which fields can be localized without breaking an FDSN query? What can a reviewer say when an event has more than one origin or magnitude? How are changes documented when an initial automatic solution is replaced by a reviewed one?

These are not isolated vocabulary puzzles. They are evidence-identity puzzles. A scientific record links a claim to a time, an instrument, a coordinate frame, a data product, an algorithm, an agency and a version. If one of those links changes, the translated statement may lead a downstream reader to the wrong waveform, an incorrect unit, a different event, or a degree of certainty that no scientist assigned. The translator’s job is to preserve the chain while making its meaning intelligible.

This guide deliberately does not replace the existing narrower earthquake bulletin, magnitude-scale and ShakeMap translation tutorial. That article concentrates on public-facing earthquake communication. This distinct master chapter concentrates on network instrumentation, archival waveform metadata, FDSN services, event association, measurement provenance and reproducible scientific records, with public bulletins discussed only as a boundary condition.

The hidden problem: words are attached to changing evidence

Seismic records have at least five simultaneously active layers. Physical ground motion occurs in the world; sensors convert a physical response into voltages or digital counts; archive formats preserve sampled time series and metadata; event processing associates arrivals and amplitudes into one or more candidate earthquake solutions; publications summarize selected solutions for an audience. A translator who enters at the publication layer but assumes that every upstream field is settled may incorrectly collapse the many possible interpretations into one definitive sentence. The visible label is only the surface of a structured chain.

Consider the fictional note, ‘origin 03:14:08 UTC; depth 12 km; evaluation mode automatic; preferred magnitude 4.8 ML; updated 03:18 UTC’. In ordinary prose it may become ‘A magnitude 4.8 earthquake occurred at 03:14 and was verified at 03:18’. The first portion describes an estimated event origin; the second has smuggled in verification. An update time alone says nothing about whether a human reviewed the solution. ML identifies a particular magnitude type; its omission invites readers to assume generic moment magnitude. Even an otherwise elegant translation can invent a certification process that did not occur.

Reproducibility raises the stakes. A researcher needs to retrieve the same source traces, apply the same instrument response, know the same epoch and compare the same event revision. Localizing an archive identifier, swapping coordinate order, omitting a channel location code or silently rounding timestamps may cause a software query to return different records. Because both source and target sentences can remain grammatically natural, normal proofreading will not catch this class of error. A technical revision must inspect the data model, not just the sentence.

Terminology also shifts with the observing system. A word such as ‘station’ may refer to a facility, a station code inside a network, a deployed instrument, or an active time period. ‘Origin’ may refer to a fitted event origin, not where a data file began. ‘Amplitude’ is measured in the context of a specific phase, channel, filter and unit, not an unconditional severity index. ‘Reviewed’ describes an evaluation state, not a guarantee of zero uncertainty. Every ambiguous noun needs a source-controlled referent before translation.

An additional difficulty is that public earthquake language and technical seismology often share vocabulary but not operational precision. ‘A tremor was felt’ describes human experience. A calibrated displacement trace describes an instrumented estimate. Modified Mercalli intensity, earthquake magnitude, peak ground acceleration and a moment tensor answer different questions; no translation should merge them into a single scale. Translation quality is therefore measured by the decisions a competent reader can still make, not by how smoothly the text reads.

Fifty mechanisms: where seismic meaning is actually stored

1. Event record versus trace

What this means. A catalogue object represents an interpreted seismic event; a waveform is a time-indexed measurement stream from a sensor. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. Confusing the two makes a catalogue estimate appear to be directly observed on one instrument. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Check whether the source is an event service result, a waveform file or explanatory prose. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Translate ‘event solution’ as an estimate; retain channel-level descriptions for recordings. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

2. Network identifier

What this means. A network code identifies the operator’s namespace rather than the geographic country or the earthquake itself. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. An invented expansion can point software and citations to a different data owner. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Consult registered network metadata and preserve the code literally. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Explain ownership separately if the publication needs it. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

3. Station code

What this means. A station label is an archive identifier with an operating history, not automatically a town name. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. Transliterating or expanding it may break a waveform query or imply a false station location. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Check network, station code and valid station epoch together. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Keep the raw label unchanged in references and machine exports. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

4. Location code

What this means. An empty location code can be meaningful and is not equivalent to an arbitrary missing value. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. Replacing the empty string with ’00’ selects another group of channels or fails the query. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Check the network’s documented location-code semantics. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Display ’empty location code’ only in reader-facing notes, not inside the identifier. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

5. Band code

What this means. FDSN channel band conveys sampling-rate range and approximate response band, not a translated adjective. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. Rendering B as ‘broad’ inside a machine key alters the channel identity. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Use the published FDSN code definitions for context. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Keep the symbolic band as a code and explain it nearby. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

6. Source code

What this means. The channel source element identifies the instrument or observed quantity within the FDSN convention. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. Confusing seismic velocity, acceleration and other sensor types changes the physical meaning. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Verify code definitions and physical units in StationXML. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Do not infer a measurement solely from casual wording. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

7. Subsource orientation

What this means. Z, N, E, 1 or 2 can encode component orientation under specified conventions. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. Assuming every horizontal axis is exactly north or east can corrupt directional analysis. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Consult channel azimuth, dip and source-code guidance. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Translate the meaning without rewriting the subsource code. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

8. Source Identifier

What this means. The URI-like FDSN source identifier combines network, station, location, band, source and subsource. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. Losing an underscore, empty code or namespace destroys a stable reference. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Compare the exact identifier with its provenance and corresponding legacy NSLC fields. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Keep the full string byte-for-byte. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

9. Station epoch

What this means. A station may be moved, recalibrated or re-equipped while retaining the same label. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. An undated station description can be applied to the wrong waveform interval. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Compare the trace date with StationXML start and end dates. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. State the valid epoch when interpreting a response. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

10. Instrument response

What this means. The response model maps recording values through the sensor and digitizer to physical units. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. Calling raw digital counts physical displacement invents calibration. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Check stages, units, gains and response validity. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Preserve whether data are uncorrected or response-corrected. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

11. Instrument sensitivity

What this means. Sensitivity connects input and output under stated conditions; it is not a generic quality score. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. Confusing sensitivity with noise floor or precision mischaracterizes data quality. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Read response metadata and associated frequency conventions. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Explain it as a documented instrument relationship. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

12. Digital counts

What this means. Counts are quantized digitizer outputs, not necessarily metres per second or metres per second squared. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. A translator who relabels counts as acceleration changes the numerical claim. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Locate the input/output units and any processing history. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Use counts unless a documented response removal establishes another unit. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

13. Sample rate

What this means. Samples per second determine time spacing; they are not a magnitude or seismic frequency of an event. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. Dropping the rate or converting it to seconds inaccurately changes the record. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Read header sample-rate/period representation and data duration. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Retain units such as samples/s explicitly. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

14. Record start time

What this means. Waveform start time marks the first sample or record boundary; it is not an earthquake origin time. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. A narrative may make the event begin when the recorder started. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Distinguish recorder time, trigger time, pick time and estimated origin. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Label each time by its role. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

15. UTC and offsets

What this means. Seismic exchange ordinarily depends on unambiguous times across networks and continents. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. Treating a UTC timestamp as local civil time shifts phase arrivals and event associations. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Retain a precise UTC designator and verify parsing. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Offer local display times only as a derived view. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

16. Clock quality

What this means. A waveform can contain timing uncertainty even when its timestamp has millisecond precision. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. Mistaking digit count for time confidence overstates arrival accuracy. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Inspect quality flags, clock-correction and instrument status where available. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Report uncertainty rather than manufacturing precision. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

17. Data gap

What this means. An unrecorded interval is not evidence that the ground was motionless. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. A translation of ‘gap’ as ‘quiet period’ can invalidate non-detection conclusions. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Inspect availability metadata and timing discontinuities. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Call it missing or unavailable observations. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

18. Data overlap

What this means. Two records may cover the same interval because of reprocessing or duplicate delivery. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. Calling overlap an extra earthquake creates a fictitious event. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Compare record provenance, sequence and sample alignment. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Preserve the duplication issue without claiming new motion. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

19. Clipping

What this means. Sensor or digitizer saturation limits the measured amplitude; a flat-topped signal is not necessarily a plateau in ground motion. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. Using the clipped maximum as the event peak can bias amplitude and magnitude. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Check instrument range and waveform shape. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Describe the measurement limitation explicitly. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

20. Background noise

What this means. Traffic, weather, machinery and electronic noise may affect recordings without representing an earthquake. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. Changing ‘coherent noise’ to ‘seismic activity’ misstates the source. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Compare channels, stations and frequency content. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Keep a noise label provisional when source attribution is unresolved. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

21. Phase pick

What this means. A phase pick marks an analyst or algorithm estimate of an arrival, not necessarily a confirmed physical onset. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. Translating every pick as ‘detected earthquake’ changes an observation into an event claim. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Check pick time, phaseHint, uncertainty and evaluation mode. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Keep arrival and event status separate. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

22. P-wave

What this means. A P-phase label identifies an arrival model and wave type in context, not a guarantee that the pick is correct. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. A mislabelled arrival changes travel-time fitting. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Compare phase coding and station distance. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Preserve the exact phase identifier in tables. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

23. S-wave

What this means. An S-phase may be absent, uncertain or difficult to pick on some records. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. Inventing an S arrival to complete a narrative affects distance and origin calculations. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Verify actual pick and uncertainty fields. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Use ‘not identified’ only when evidence says so. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

24. Arrival residual

What this means. Residuals compare observed and predicted arrival times under an origin and travel-time model. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. Misreading residual as total travel time changes diagnostics. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Check sign convention, units and method context. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Translate as a model discrepancy, not event duration. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

25. Origin estimate

What this means. An origin is a fitted set of location and time parameters associated with an event hypothesis. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. Calling it the unique true epicentre hides competing solutions. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Inspect preferred origin and evaluation stage. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Use estimated origin when uncertainty remains. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

26. Epicentre and hypocentre

What this means. Epicentre is a surface projection; hypocentre or focus includes depth. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. Using both interchangeably removes the vertical location dimension. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Check horizontal coordinates and depth reference. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Retain the distinction in scientific summaries. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

27. Depth estimate

What this means. Depth may be fixed, constrained or freely estimated, depending on the solution method. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. Rendering ‘fixed depth’ as a confident measured depth misstates the inference. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Check depth type and method-specific metadata. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Say fixed for processing when that is what happened. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

28. Horizontal uncertainty

What this means. A location error ellipse or covariance captures uncertainty, not a literal boundary of the fault rupture. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. Translating the ellipse as the event’s physical footprint confuses statistics and geology. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Inspect confidence level, axes and units. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Tie uncertainty to the fitted location. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

29. Vertical uncertainty

What this means. Depth uncertainty can differ radically from horizontal error even for one solution. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. Dropping vertical uncertainty suggests the 3-D position is equally reliable. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Look for separate depth uncertainty and fixed-depth flags. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Do not infer precision from decimal places. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

30. Event association

What this means. Multiple phase picks are grouped into a hypothesized event solution by processing rules. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. A wrong association can make two separate occurrences appear to be one. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Compare event identifiers, origin links and processing histories. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Use event association as a relational term. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

31. Multiple event IDs

What this means. One physical event can appear under several agency identifiers and later be merged or cross-referenced. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. Treating each code as a separate earthquake inflates counts. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Inspect cross-reference IDs, contributors and preferred products. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Retain agency keys and record deduplication decisions. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

32. Magnitude type

What this means. ML, mb, Ms and Mw are not unqualified synonyms or automatically interchangeable scales. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. Replacing a specific type with another changes the measurement model. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Check magType alongside the numeric estimate. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Keep type and uncertainty visible. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

33. Magnitude uncertainty

What this means. The number of displayed decimals does not define the confidence of a magnitude estimate. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. Promoting 4.63 to ‘exactly 4.63’ suppresses uncertainty. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Read uncertainty fields and the agency’s rounding policy. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Preserve qualifying language. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

34. Moment tensor

What this means. A focal or moment-tensor solution describes aspects of the source model, not a direct photograph of a fault. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. Translating nodal planes as proven ruptured planes makes an unwarranted claim. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Review the inversion method and its ambiguity. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Describe alternative fault-plane interpretations where appropriate. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

35. Evaluation mode

What this means. Automatic and manual or reviewed processing statuses reflect how a solution was produced. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. An updated automatic solution is not automatically human-verified. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Inspect evaluationMode and evaluationStatus separately. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Do not invent independent review from publication time. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

36. Preferred origin

What this means. A data provider may nominate one solution while preserving alternatives in the same event object. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. A translation listing all origins as simultaneous earthquakes misreads the structure. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Check preferredOriginID and origin references. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Mark the preferred value and source. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

37. QuakeML publicID

What this means. A publicID links structured objects and is not a prose label meant for localization. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. Translating any part of an ID breaks cross references. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Validate ID strings and object links after export. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Leave IDs exact and annotate meaning outside them. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

38. StationXML metadata

What this means. StationXML describes networks, stations, channels and instrument responses needed to interpret data. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. Treating it as a waveform file conceals its metadata function. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Compare StationXML epochs and response chain with a trace. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Translate descriptions, not XML element names or identifiers. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

39. miniSEED waveform archive

What this means. miniSEED records store sampled time series and associated headers; version 3 has defined new header features. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. Changing binary payloads to localize captions risks corrupting the scientific data. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Keep file bytes unchanged and localize surrounding documentation. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Document whether the archive is miniSEED 2 or 3. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

40. FDSN station service

What this means. fdsnws-station supplies metadata rather than the sampled waveform itself. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. A URL described as downloading measured signal might only yield the station description. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Check service endpoint, query parameters and output type. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Differentiate metadata retrieval from data retrieval. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

41. FDSN dataselect service

What this means. fdsnws-dataselect retrieves time series based on source selection and time intervals. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. Translating a query key or dropping the window changes which samples are retrieved. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Preserve API parameter names, codes and timestamps. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Demonstrate queries as inert examples, not live alerts. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

42. FDSN event service

What this means. fdsnws-event exchanges event parameters, often through QuakeML representation. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. An event listing is not an archive of all station samples. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Inspect service response type and version. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Keep the distinction between model parameters and raw observations. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

43. FDSN availability service

What this means. Availability gives an account of accessible waveform intervals; it does not promise continuous usable signals. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. Calling a span ‘uninterrupted seismic coverage’ may hide gaps or quality limitations. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Read interval and extent semantics. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Treat absence of access separately from absence of earthquakes. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

44. SeedLink streaming

What this means. Streaming protocols distribute waveform packets and have different timing/replay characteristics from archive queries. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. A delivered packet is not evidence of a processed or reviewed event. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Check stream identifiers, sequence and latency. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Describe stream state without inflating scientific certainty. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

45. Historical bulletin

What this means. Legacy printed phase readings are archival observations with historical instrumentation and conventions. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. Modern terminology may imply precision or calibration that was not available. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Preserve original notation with an editorial explanation. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Separate transcription uncertainty from geophysical uncertainty. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

46. Release and revision history

What this means. Earthquake products may be regenerated, superseded or associated with new preferred solutions. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. A translation that drops version and update time makes results irreproducible. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Compare generation time, identifier, source and update log. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. State whether values are current at a named retrieval. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

47. Intensity versus magnitude

What this means. Magnitude estimates source size; intensity describes shaking effects at a place by a particular intensity method. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. A numerical value from one cannot be relabelled as the other. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Check scale name, geographic locality and provenance. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Keep source-size and experienced shaking distinct. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

48. Amplitude and filter

What this means. Amplitude depends on instrument, wave phase, component, units and filtering procedure. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. Comparing filtered amplitudes as if they were identical raw measurements can mislead. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Inspect processing chain and frequency band. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Describe amplitude with the necessary qualifiers. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

49. Waveform provenance

What this means. Data may pass through instruments, recorders, repositories and scientific processing tools. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. An unlabeled derivative can be mistaken for untouched original samples. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Record raw and derived versions with transformations. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Keep source DOIs or persistent records when provided. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

50. Data rights and access

What this means. Open metadata, restricted waveform access and licensed processing products are distinct permissions. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. ‘Available’ does not always mean publicly reusable. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Inspect access flags, licenses and repository instructions. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Translate access status as a legal-operational field. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

51. Cross-network reproducibility

What this means. Networks may differ in naming, thresholds, instrumentation and review practices. In seismic translation, the correct unit of meaning is not simply the nearest dictionary equivalent. It is the definition as applied to a record, a method and an intended reader. Ask which entity is being described, which attributes belong to it, and which values are controlled vocabulary rather than natural-language expressions.

Hidden failure. Merging records without these differences creates false comparability. This is important because a target sentence may sound normal while silently altering the field’s object, certainty or units. The result can mislead an analyst and also break a query, comparison, citation or data conversion. A fluent sentence is therefore not sufficient evidence of faithful transmission.

Repeatable check. Document agency, method, epoch and data coverage. For this mechanism the translator should build a small ledger with original field name, exact value, human-readable gloss, provenance source and reviewer decision. The bilingual reviewer tests whether the same record can still be found and whether the same scientific conclusion can still be defended. Use the same technical explanation standard across languages. Where the source lacks enough information, preserve the ambiguity and ask the data owner rather than adding a plausible fact.

The repeatable method: eighteen release gates

Gate 01: Identify the deliverable

Decision. Decide whether this work is an archive handover, a database interface, a scientific report or a public explanation. At this point the team should write down who is allowed to approve the decision, what proof is sufficient and what remains out of scope. A meaningful quality gate has a possible failure condition; it is not a box to tick because a sentence reads pleasantly.

Test. The wrong document type invites unjustified simplification. Record the audience, data format, purpose and expected downstream use. Build one positive test using known data and one negative test designed to catch the stated error. Store both the observation and the reviewer decision in a compact ledger. If the underlying evidence cannot be inspected, mark the relevant claim unresolved, explain the limitation plainly and avoid release where the misunderstanding could change a scientific or operational decision.

Gate 02: Freeze the source snapshot

Decision. Retain the exact extracted text and a stable copy of relevant data or documented retrieval time. At this point the team should write down who is allowed to approve the decision, what proof is sufficient and what remains out of scope. A meaningful quality gate has a possible failure condition; it is not a box to tick because a sentence reads pleasantly.

Test. Without a snapshot, a changing earthquake catalogue can shift under the translator’s feet. Record event IDs, query windows, product versions and access date. Build one positive test using known data and one negative test designed to catch the stated error. Store both the observation and the reviewer decision in a compact ledger. If the underlying evidence cannot be inspected, mark the relevant claim unresolved, explain the limitation plainly and avoid release where the misunderstanding could change a scientific or operational decision.

Gate 03: Separate layers

Decision. Classify each item as data value, identifier, metadata label, scientific interpretation or public-language message. At this point the team should write down who is allowed to approve the decision, what proof is sufficient and what remains out of scope. A meaningful quality gate has a possible failure condition; it is not a box to tick because a sentence reads pleasantly.

Test. A label may be translatable while its value is not. Mark machine-controlled strings as immutable. Build one positive test using known data and one negative test designed to catch the stated error. Store both the observation and the reviewer decision in a compact ledger. If the underlying evidence cannot be inspected, mark the relevant claim unresolved, explain the limitation plainly and avoid release where the misunderstanding could change a scientific or operational decision.

Gate 04: Map object relationships

Decision. Draw a small diagram linking event, origins, magnitudes, picks, stations, channels and data products. At this point the team should write down who is allowed to approve the decision, what proof is sufficient and what remains out of scope. A meaningful quality gate has a possible failure condition; it is not a box to tick because a sentence reads pleasantly.

Test. An elegant paragraph can detach an observation from its actual event. Verify every displayed reference against its linked object. Build one positive test using known data and one negative test designed to catch the stated error. Store both the observation and the reviewer decision in a compact ledger. If the underlying evidence cannot be inspected, mark the relevant claim unresolved, explain the limitation plainly and avoid release where the misunderstanding could change a scientific or operational decision.

Gate 05: Specify the language function

Decision. Decide whether target text must enable reading, searching, scientific reproduction or public understanding. At this point the team should write down who is allowed to approve the decision, what proof is sufficient and what remains out of scope. A meaningful quality gate has a possible failure condition; it is not a box to tick because a sentence reads pleasantly.

Test. One rendering may be good prose but bad data documentation. State the permitted explanatory additions. Build one positive test using known data and one negative test designed to catch the stated error. Store both the observation and the reviewer decision in a compact ledger. If the underlying evidence cannot be inspected, mark the relevant claim unresolved, explain the limitation plainly and avoid release where the misunderstanding could change a scientific or operational decision.

Gate 06: Establish unit policy

Decision. Lock all physical units, their reference quantities and any approved conversion procedure. At this point the team should write down who is allowed to approve the decision, what proof is sufficient and what remains out of scope. A meaningful quality gate has a possible failure condition; it is not a box to tick because a sentence reads pleasantly.

Test. Changing counts into acceleration without processing is false measurement. Retain units with values and note derived-unit provenance. Build one positive test using known data and one negative test designed to catch the stated error. Store both the observation and the reviewer decision in a compact ledger. If the underlying evidence cannot be inspected, mark the relevant claim unresolved, explain the limitation plainly and avoid release where the misunderstanding could change a scientific or operational decision.

Gate 07: Establish time policy

Decision. Lock UTC, precision, origin time, pick time, recorder time, modification time and epoch validity. At this point the team should write down who is allowed to approve the decision, what proof is sufficient and what remains out of scope. A meaningful quality gate has a possible failure condition; it is not a box to tick because a sentence reads pleasantly.

Test. A timezone shift can destroy cross-station timing analysis. Use unambiguous timestamps with labeled roles. Build one positive test using known data and one negative test designed to catch the stated error. Store both the observation and the reviewer decision in a compact ledger. If the underlying evidence cannot be inspected, mark the relevant claim unresolved, explain the limitation plainly and avoid release where the misunderstanding could change a scientific or operational decision.

Gate 08: Establish identifier policy

Decision. Lock FDSN source IDs, network codes, station codes, publicIDs, URLs and object references. At this point the team should write down who is allowed to approve the decision, what proof is sufficient and what remains out of scope. A meaningful quality gate has a possible failure condition; it is not a box to tick because a sentence reads pleasantly.

Test. Localizing characters inside codes breaks machine and human cross references. Preserve exact bytes and attach separate readable explanations. Build one positive test using known data and one negative test designed to catch the stated error. Store both the observation and the reviewer decision in a compact ledger. If the underlying evidence cannot be inspected, mark the relevant claim unresolved, explain the limitation plainly and avoid release where the misunderstanding could change a scientific or operational decision.

Gate 09: Trace station metadata

Decision. Associate each example waveform with the station and channel epoch valid at its timestamp. At this point the team should write down who is allowed to approve the decision, what proof is sufficient and what remains out of scope. A meaningful quality gate has a possible failure condition; it is not a box to tick because a sentence reads pleasantly.

Test. Modern response metadata applied to older traces can change derived motion. Save the checked StationXML epoch. Build one positive test using known data and one negative test designed to catch the stated error. Store both the observation and the reviewer decision in a compact ledger. If the underlying evidence cannot be inspected, mark the relevant claim unresolved, explain the limitation plainly and avoid release where the misunderstanding could change a scientific or operational decision.

Gate 10: Validate waveform interpretation

Decision. Check sample rate, sensor response, raw or processed state, clipping, gaps, overlaps and timing quality. At this point the team should write down who is allowed to approve the decision, what proof is sufficient and what remains out of scope. A meaningful quality gate has a possible failure condition; it is not a box to tick because a sentence reads pleasantly.

Test. Missing samples or saturating signals cannot become confident negative or amplitude evidence. Write limitations visibly. Build one positive test using known data and one negative test designed to catch the stated error. Store both the observation and the reviewer decision in a compact ledger. If the underlying evidence cannot be inspected, mark the relevant claim unresolved, explain the limitation plainly and avoid release where the misunderstanding could change a scientific or operational decision.

Gate 11: Validate event association

Decision. Check whether picks, origins and magnitudes belong to the same event hypothesis and revision. At this point the team should write down who is allowed to approve the decision, what proof is sufficient and what remains out of scope. A meaningful quality gate has a possible failure condition; it is not a box to tick because a sentence reads pleasantly.

Test. A cross-link error can create a phantom earthquake or merge two events. Use object IDs and provenance instead of prose similarity. Build one positive test using known data and one negative test designed to catch the stated error. Store both the observation and the reviewer decision in a compact ledger. If the underlying evidence cannot be inspected, mark the relevant claim unresolved, explain the limitation plainly and avoid release where the misunderstanding could change a scientific or operational decision.

Gate 12: Respect evaluation status

Decision. Carry automatic, preliminary, reviewed, rejected and preferred distinctions only as supported by data. At this point the team should write down who is allowed to approve the decision, what proof is sufficient and what remains out of scope. A meaningful quality gate has a possible failure condition; it is not a box to tick because a sentence reads pleasantly.

Test. Publication and review are separate processes. Never infer validation from an updated timestamp. Build one positive test using known data and one negative test designed to catch the stated error. Store both the observation and the reviewer decision in a compact ledger. If the underlying evidence cannot be inspected, mark the relevant claim unresolved, explain the limitation plainly and avoid release where the misunderstanding could change a scientific or operational decision.

Gate 13: Draft by controlled concepts

Decision. Write the target meaning around stable objects, quantities, statuses and relationships. At this point the team should write down who is allowed to approve the decision, what proof is sufficient and what remains out of scope. A meaningful quality gate has a possible failure condition; it is not a box to tick because a sentence reads pleasantly.

Test. An early free rewrite may embed an irreversible assumption. Use short clauses with labeled evidence. Build one positive test using known data and one negative test designed to catch the stated error. Store both the observation and the reviewer decision in a compact ledger. If the underlying evidence cannot be inspected, mark the relevant claim unresolved, explain the limitation plainly and avoid release where the misunderstanding could change a scientific or operational decision.

Gate 14: Run numerical and code QA

Decision. Compare timestamps, numeric fields, coordinate signs, scientific notation, decimal precision and protected strings. At this point the team should write down who is allowed to approve the decision, what proof is sufficient and what remains out of scope. A meaningful quality gate has a possible failure condition; it is not a box to tick because a sentence reads pleasantly.

Test. Small formatting changes can cause large geophysical changes. Run field-by-field diffs and sample retrieval checks. Build one positive test using known data and one negative test designed to catch the stated error. Store both the observation and the reviewer decision in a compact ledger. If the underlying evidence cannot be inspected, mark the relevant claim unresolved, explain the limitation plainly and avoid release where the misunderstanding could change a scientific or operational decision.

Gate 15: Run bilingual science QA

Decision. Ask a domain reviewer to compare the translated claim with the source event model and waveform context. At this point the team should write down who is allowed to approve the decision, what proof is sufficient and what remains out of scope. A meaningful quality gate has a possible failure condition; it is not a box to tick because a sentence reads pleasantly.

Test. A translator may preserve grammar while changing causal interpretation. Treat disagreement as a concept issue, not merely style. Build one positive test using known data and one negative test designed to catch the stated error. Store both the observation and the reviewer decision in a compact ledger. If the underlying evidence cannot be inspected, mark the relevant claim unresolved, explain the limitation plainly and avoid release where the misunderstanding could change a scientific or operational decision.

Gate 16: Test readers by role

Decision. Give a scientist, archivist and non-specialist a task appropriate to their use case. At this point the team should write down who is allowed to approve the decision, what proof is sufficient and what remains out of scope. A meaningful quality gate has a possible failure condition; it is not a box to tick because a sentence reads pleasantly.

Test. Different audiences expose different hidden errors. A scientist should locate the same product; a lay reader should not overread certainty. Build one positive test using known data and one negative test designed to catch the stated error. Store both the observation and the reviewer decision in a compact ledger. If the underlying evidence cannot be inspected, mark the relevant claim unresolved, explain the limitation plainly and avoid release where the misunderstanding could change a scientific or operational decision.

Gate 17: Document unresolved issues

Decision. Record query decisions, unknown values, conflicting agency entries and rejected alternatives. At this point the team should write down who is allowed to approve the decision, what proof is sufficient and what remains out of scope. A meaningful quality gate has a possible failure condition; it is not a box to tick because a sentence reads pleasantly.

Test. Silent guessing makes later repairs hard to trace. Maintain a change ledger with accountable approvals. Build one positive test using known data and one negative test designed to catch the stated error. Store both the observation and the reviewer decision in a compact ledger. If the underlying evidence cannot be inspected, mark the relevant claim unresolved, explain the limitation plainly and avoid release where the misunderstanding could change a scientific or operational decision.

Gate 18: Release, monitor and correct

Decision. Publish only after the required checks and retain the source snapshot and correction procedure. At this point the team should write down who is allowed to approve the decision, what proof is sufficient and what remains out of scope. A meaningful quality gate has a possible failure condition; it is not a box to tick because a sentence reads pleasantly.

Test. A later catalogue revision can legitimately supersede earlier wording. Label versions, preserve audit history and route corrections through controlled review. Build one positive test using known data and one negative test designed to catch the stated error. Store both the observation and the reviewer decision in a compact ledger. If the underlying evidence cannot be inspected, mark the relevant claim unresolved, explain the limitation plainly and avoid release where the misunderstanding could change a scientific or operational decision.

Failure modes: what can go wrong even after proofreading

Failure 1: A station code becomes a city name

Changed claim: A code does not prove a geographic expansion; copied prose may target the wrong sensor. Repair: Compare the full registered identifier and network metadata. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 2: An empty location code becomes 00

Changed claim: These can select different channel groups and distinct instruments. Repair: Retest the waveform query using an exact location field. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 3: A legacy BHZ code becomes an English phrase inside a URL

Changed claim: Descriptive translation is not allowed inside the machine path. Repair: Keep the code intact and explain it outside the URL. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 4: A channel epoch is lost

Changed claim: The selected calibration may not apply to the waveform’s timestamp. Repair: Rejoin time-indexed channel metadata. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 5: Raw counts become acceleration

Changed claim: A response correction was never performed or documented. Repair: Trace sensor units and processing stage. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 6: A response gain is called accuracy

Changed claim: Gain relates a physical input to an output; accuracy is another concept. Repair: Review the response model and intended wording. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 7: A sampling rate becomes a time interval

Changed claim: Samples per second and seconds per sample are reciprocals. Repair: Validate arithmetic and units explicitly. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 8: UTC disappears from a timestamp

Changed claim: A reader may interpret the clock time in another timezone. Repair: Keep the UTC or offset designation. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 9: A recorder start becomes the earthquake origin

Changed claim: Recording coverage and source-origin estimates are different timestamps. Repair: Check event origin against waveform header. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 10: A pick time becomes a detection time

Changed claim: An arrival estimate at one station is not the event’s existence proof. Repair: Preserve the pick object and phase association. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 11: A P pick is treated as certain

Changed claim: Phase labels may be provisional or have uncertainties. Repair: Carry evaluation mode and pick errors. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 12: A missing S pick is filled in

Changed claim: Blank evidence becomes invented measurement. Repair: Mark not available only when source supports it. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 13: A gap is translated as calm ground

Changed claim: Absent samples do not show zero motion. Repair: Read waveform availability and quality. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 14: A saturated waveform yields a quoted peak

Changed claim: Clipped values may not represent actual extrema. Repair: Flag clipping and reject unjustified peak inference. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 15: Duplicate time records become two events

Changed claim: Archive overlap is not seismic repetition. Repair: Check trace versions and event IDs separately. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 16: A fixed depth is presented as measured

Changed claim: An algorithm constraint can be mistaken for observation. Repair: Carry the fixed-depth qualifier. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 17: Epicentre and hypocentre are merged

Changed claim: Surface coordinates cannot substitute for depth and focus. Repair: Keep both physical definitions. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 18: Coordinate latitude and longitude switch

Changed claim: A valid-looking coordinate becomes a different geographic point. Repair: Check sign, order and spatial convention. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 19: Depth error is assumed equal to horizontal error

Changed claim: Uncertainty is multidimensional and method-dependent. Repair: Review each reported uncertainty separately. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 20: A first automatic magnitude is called final

Changed claim: A preliminary solution is incorrectly given authority. Repair: Inspect both method and evaluation status. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 21: ML becomes Mw

Changed claim: Different magnitude methods are replaced with a generic symbol. Repair: Preserve magType and compare method definitions. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 22: Intensity is translated as magnitude

Changed claim: Local shaking or damage observations become an event-size estimate. Repair: Keep scale and measure distinct. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 23: An event update becomes a new earthquake

Changed claim: Version identifiers are mistaken for unique physical events. Repair: Inspect associations and product histories. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 24: Two agencies’ IDs are deduplicated by title alone

Changed claim: Similar location words are not proof of identical events. Repair: Use cross references, origin times and provenance. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 25: A preferred origin is called the only origin

Changed claim: Alternative solutions vanish from the translated record. Repair: Keep preference distinct from uniqueness. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 26: QuakeML publicID is localized

Changed claim: Structural cross references break when characters change. Repair: Perform byte-for-byte validation. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 27: A StationXML attribute is renamed

Changed claim: The output no longer satisfies the original schema. Repair: Translate documentation text instead of keys. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 28: miniSEED 2 and 3 are treated as interchangeable bytes

Changed claim: Format-version differences are suppressed. Repair: Record exact format version and parser expectations. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 29: A waveform file is called an event catalogue

Changed claim: Time series and model summaries get conflated. Repair: Identify the actual output format before translation. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 30: Data availability is called non-occurrence

Changed claim: No accessible record is not evidence of no earthquake. Repair: Report access limits honestly. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 31: Stream latency disappears

Changed claim: A reader assumes real-time information was instantaneous. Repair: Carry timestamp and collection or delivery delay. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 32: A local time conversion is undocumented

Changed claim: Apparent event order changes in comparison tables. Repair: Keep UTC master columns and a labeled derived view. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 33: Noise is treated as a verified quake

Changed claim: Instrument disturbances may not be seismic events. Repair: Mark interpretation as provisional until supported. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 34: A filtered trace is called raw

Changed claim: Signal-processing transformations are hidden. Repair: Record processing steps and retained original. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 35: Displacement and velocity share one label

Changed claim: Distinct derivatives of motion are merged. Repair: Keep m, m/s and m/s² linked to correct variables. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 36: A quoted significance score becomes a magnitude

Changed claim: Product ranking or significance is not a magnitude estimate. Repair: Read the provider-specific field definition. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 37: An uncertainty value becomes a prediction interval

Changed claim: Different statistical meanings are merged. Repair: Verify the underlying confidence convention. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 38: Historical station notation is modernized silently

Changed claim: Editorial normalization overwrites the evidence trail. Repair: Keep diplomatic transcription plus labeled clarification. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 39: A data DOI is stripped

Changed claim: The reader cannot find or cite the original observation. Repair: Preserve persistent citation identifiers. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 40: A restricted dataset is called open

Changed claim: Access and reuse rights are confused. Repair: Inspect the documented license and access conditions. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 41: A human review is inferred from the word updated

Changed claim: Update is a clock event, not a quality certification. Repair: Inspect evaluation status and human-review records. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 42: A parameter is rounded for appearance

Changed claim: Small rounding may change threshold grouping or exact reproducibility. Repair: Keep original precision and a clear display policy. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Failure 43: A translation reads like an instruction to evacuate

Changed claim: A scientific scenario has accidentally become a public warning. Repair: Separate illustrative analysis from authority-issued directions. The reviewer should locate the earliest field where the mistaken interpretation entered the target text and document its downstream effects. Do not merely replace the suspicious word; verify numeric neighbors, linked records, format labels and source chronology.

Use a two-column check. On the left put the literal source field, its value and the cited artifact. On the right put the candidate translation and the decision it would cause a reader to make. If that decision differs from what the source permits, revise or explicitly qualify the target sentence. This habit catches semantic errors that ordinary spelling and fluency tools routinely miss.

Worked diagnostics: 84 fictional records and repair decisions

Diagnostic 01: Network namespace

Fictional source situation. Event network XY supplies station ABCD, not a geographic abbreviation. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The network code is expanded to an invented country. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Keep XY as network ID, with ownership described only if separately verified. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check FDSN network registry and station metadata. A query or citation could target an unrelated operator. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 02: Station alias

Fictional source situation. A survey narrative calls station Q12 ‘ridge hut’ while the archive uses Q12. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The human nickname replaces Q12 in the data export. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Retain the original station identifier and place the nickname in commentary. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check StationXML station code and station description. Future data retrieval will fail or diverge. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 03: Empty location

Fictional source situation. Trace XY.Q12..BHZ uses an empty legacy location code. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The double separator is silently changed to 00. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Preserve the empty location; explain it in text if necessary. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Source Identifier mapping rules and archive query. The wrong instrument stream may be selected. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 04: Three-component comparison

Fictional source situation. BHZ, BHN and BHE are listed as separate channels for one epoch. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The three channels are described as three earthquake events. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Treat them as components of a station’s observations. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Channel metadata and shared event association. The event count becomes fictitious. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 05: Rotated horizontal channels

Fictional source situation. The deployment labels its horizontal components BH1 and BH2. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The translation forces the axes to north and east. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Retain 1 and 2 with recorded azimuth and dip. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check StationXML orientation parameters. Incorrectly rotated traces distort directional research. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 06: High-rate channel

Fictional source situation. A channel coded EHZ is sampled at 100 samples per second. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The letter E is translated as an eastward motion indicator. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Distinguish band from subsource and preserve the code. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check FDSN channel code definitions. Sensor type and filtering may be misread. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 07: Location group 10

Fictional source situation. Station ST01 has location 00 and 10 from separate digitizers. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The publication drops location 10 as unnecessary notation. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Keep the location field in every exported identifier. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Network’s channel grouping documentation. Independent streams can be merged accidentally. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 08: Station relocation

Fictional source situation. Station ST05 moved after a year boundary but reused its code. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The old waveform is assigned the new coordinates. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Join by both station code and epoch date. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Historical StationXML epochs. Researchers compute incorrect station distances. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 09: Sensor replacement

Fictional source situation. A broadband instrument was exchanged during a maintenance interval. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. A single current sensitivity value is attached to all years. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Use the response belonging to the waveform date. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Response epoch and equipment logs. Amplitudes may be biased after correction. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 10: Digitizer gain change

Fictional source situation. A recorder changed gain without changing the public station title. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The translator calls old raw counts identical physical units. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Report the gain epochs, not one timeless description. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Channel response stages and change log. Historical measurements become falsely comparable. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 11: Counts only

Fictional source situation. An archived trace is labeled ‘uncorrected counts’. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. A public rewrite calls the ordinate ground velocity in m/s. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Keep counts and explain that calibration is separate. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Original data header and documented processing. A scientifically false unit appears on the graph. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 12: Response-corrected trace

Fictional source situation. A processed trace records velocity in m/s after response removal. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The target caption says ‘original raw sensor reading’. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Label it derived ground velocity with processing provenance. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Processing manifest and units. The reader may double-correct the signal. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 13: Acceleration units

Fictional source situation. A strong-motion trace is expressed in m/s². This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The translation converts the numeric values into g without calculation. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Keep m/s² or perform a documented valid conversion. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Units, conversion factor and data product description. Ground-motion comparison acquires a false scale. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 14: Displacement units

Fictional source situation. A waveform plot uses nanometres for displacement. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The translated caption shows the same numbers in millimetres. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Retain nanometres or convert with 10^6 scaling. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Plot metadata and axis annotation. Amplitude changes by a millionfold. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 15: Sampling frequency

Fictional source situation. A waveform contains 50 samples each second. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The subtitle says one observation every 50 seconds. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Write 50 samples per second, or 0.02 seconds per sample. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Sample rate and index count. Reconstructed time axis becomes wrong. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 16: Duration calculation

Fictional source situation. A trace has 6,000 samples at 100 Hz. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The translation calls it 6,000 seconds of coverage. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Calculate nominal duration from sample count and rate. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Waveform header, timing and sample count. A one-minute record becomes a fictitious long history. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 17: UTC timestamp

Fictional source situation. A phase pick at 2026-02-04 08:12:00Z is displayed without timezone. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. Readers assign their own local time zone. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Retain UTC and make any display conversion explicit. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check ISO timestamp and application timezone settings. Cross-network picks appear out of order. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 18: Time correction

Fictional source situation. A recorder log says the clock was corrected after a drift alert. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The text claims the seismic event was moved to a new date. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Distinguish instrument clock adjustment from event relocation. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Time-quality and maintenance records. The processing chronology is corrupted. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 19: Leap-second boundary

Fictional source situation. A legacy sequence crosses a leap-second handling boundary. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. A translated spreadsheet silently adds one second twice. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Keep the archive’s time representation and conversion method. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Time standard and transformation log. Arrival residuals become systematically shifted. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 20: Waveform start

Fictional source situation. A record begins 20 seconds before an estimated event origin. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The start of the trace is described as earthquake onset. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Label record coverage separately from origin and phase arrivals. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check miniSEED start time and QuakeML origin. Physical sequence is misrepresented. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 21: Data gap

Fictional source situation. A channel contains a six-second acquisition gap. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The report calls the event silent for six seconds. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Describe missing recorded data, not ground quiescence. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Availability intervals and packet flags. Unobserved motion is confused with non-motion. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 22: Packet overlap

Fictional source situation. Two files cover the same minute after a backfill. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The duplicate interval is counted twice as event duration. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Deduplicate or preserve separately with provenance. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Archive index and record timestamps. Derived amplitudes may be double-weighted. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 23: Saturation

Fictional source situation. A transient waveform clips at the digitizer limit. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The clipped height is claimed as the true maximum. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. State saturation and do not infer an unconstrained peak. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Waveform flags, response range and visual review. Magnitude or ground-motion products may be biased. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 24: Flatline electronics

Fictional source situation. A component recorded zeros after an equipment fault. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The summary says the station measured no ground motion. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Call it an instrument outage or invalid data. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check State-of-health and neighboring channels. Readers infer absence of seismic activity. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 25: Environmental noise

Fictional source situation. A strong signal aligns with local wind disturbance. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The narrative states an earthquake is confirmed by one station. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Retain the provisional non-seismic interpretation. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Cross-station coherence and sensor notes. Noise becomes an invented event. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 26: P arrival

Fictional source situation. An automatic picker tags a candidate arrival ‘P’. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The translation states an expert confirmed the first wave. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Keep candidate P-phase pick with evaluation mode. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Pick object and review flags. An algorithm suggestion becomes false authority. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 27: S arrival

Fictional source situation. A regional station has an uncertain S pick with wide error. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The report prints a precise arrival without qualifiers. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Preserve uncertainty and any analyst comment. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check QuakeML pick uncertainty and phase association. Travel-time conclusions become overconfident. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 28: Unidentified phase

Fictional source situation. A phase label is blank in an imported legacy record. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The target text supplies P because it appears first. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Keep phase unidentified pending expert assignment. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Original record and phase list. A non-evidenced phase affects association. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 29: Arrival residual

Fictional source situation. The model residual for one phase is +1.4 s. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The translation says the seismic wave lasted 1.4 s. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Describe observed minus modeled timing residual with convention. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Origin method and residual definition. The diagnostic is turned into a duration. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 30: Station distance

Fictional source situation. The station is 3.2 degrees from the event in a travel-time table. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The translator reads 3.2 kilometres. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Preserve angular degrees and label interpretation. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Distance field documentation and coordinate geometry. Travel-time method selection becomes wrong. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 31: Azimuth

Fictional source situation. A station-to-event azimuth is 280 degrees clockwise from north. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The text describes it as an elevation angle. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Retain horizontal bearing and reference frame. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Geometry metadata. An observer follows the wrong direction. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 32: Latitude sign

Fictional source situation. An event latitude is -12.5 degrees. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. A typesetter drops the minus sign. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Preserve signed latitude or explicit hemisphere consistently. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Raw coordinates and map cross-check. The reported epicentre moves hemispheres. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 33: Longitude sign

Fictional source situation. A point uses longitude -75.2 degrees. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. A localized table reverses the sign. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Keep the signed coordinate and check orientation. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Event origin and GIS validation. The event is mapped elsewhere. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 34: Coordinate order

Fictional source situation. The source record is latitude, longitude, depth. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The target summary writes longitude, latitude with same labels. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Follow source convention and verify each field label. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check QuakeML and client schema. A map may plot an unrelated point. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 35: Depth in kilometres

Fictional source situation. A source gives depth 14 km. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The target table labels the same number as metres. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Keep km or convert numerically. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Origin depth unit and rendering policy. Event depth changes by a factor of one thousand. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 36: Fixed depth

Fictional source situation. The inversion ran with depth fixed at 10 km. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The translation says scientists measured exactly 10 km. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. State that the solution constrained depth to 10 km. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Origin method and fixed-depth metadata. A methodological choice appears observational. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 37: Depth uncertainty

Fictional source situation. Horizontal error is small but depth uncertainty is large. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The target uses one generic plus-minus value for all axes. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Preserve separate position uncertainties where supplied. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Origin quality and covariance documentation. Readers overestimate depth precision. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 38: Alternative origins

Fictional source situation. Two networks propose different hypocentres for one event. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The summary calls them two earthquakes. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Show agency-specific origin hypotheses and preference. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Event association and origin IDs. Catalogue counts are inflated. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 39: Preferred origin

Fictional source situation. A provider marks one origin as preferred and retains older alternatives. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. A translation deletes the preference label and lists all equally. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Identify the preferred origin at the stated retrieval time. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check QuakeML preferredOriginID and product timestamp. A reader uses the wrong active solution. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 40: Initial magnitude

Fictional source situation. An automated estimate is ML 4.3. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The caption calls it a final authoritative Mw 4.3. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Keep preliminary ML 4.3 and agency status. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Magnitude type and evaluation mode. An unperformed magnitude conversion is implied. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 41: Magnitude revision

Fictional source situation. An event’s preferred magnitude changed from 4.2 to 4.4 after review. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The target article reports both without chronology. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Attach each value to version and update timestamp. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Catalogue product history. Changes are mistaken for two simultaneous truths. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 42: mb and Mw

Fictional source situation. A table reports mb 5.6 and Mw 5.8 from different methods. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The text says agencies disagreed about one identical scale. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Preserve labels and method distinction. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Magnitude-type definitions and producer metadata. Cross-method values are falsely treated as direct contradictions. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 43: Local magnitude

Fictional source situation. ML is used for a nearby moderate event. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The translation unqualifiedly calls it the Richter scale in all contexts. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Retain ML and explain its specific method when warranted. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Published magnitude-type mapping. A popular simplification overrides data fidelity. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 44: Magnitude decimal

Fictional source situation. An algorithm reports 3.47 with a stated uncertainty. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The target says the earthquake was exactly magnitude 3.47. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Report the estimate and uncertainty without implying exactness. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Magnitude object and method notes. Numerical display is mistaken for physical certainty. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 45: Magnitude threshold

Fictional source situation. A dataset query requests minmagnitude 4.0. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The translated guide describes values greater than 4 only. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Preserve the service’s actual inclusive/exclusive semantics. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Endpoint documentation and test query. Boundary events may be silently omitted. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 46: Event association

Fictional source situation. A cluster of arrivals is associated with candidate event EV-X1. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The target says each pick is a separate earthquake. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Keep the association and distinguish observations. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check QuakeML event and pick references. Readers overcount seismicity. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 47: Agency identifiers

Fictional source situation. Two contributing agencies reference the same occurrence with distinct IDs. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. A merged table erases the agency prefixes. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Retain each contributor ID and explicit crosswalk. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Catalogue contributor metadata and associated IDs. The chain of evidence is lost. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 48: Event type

Fictional source situation. A reviewed record labels a source ‘quarry blast’. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The translation categorizes it as an earthquake because it shook the ground. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Retain event-type classification and its authority. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Event type and evaluation history. Anthropogenic signals pollute natural earthquake counts. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 49: Uncertain event type

Fictional source situation. An automatic detector marks a ‘seismic event’ without final classification. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The target says a confirmed tectonic earthquake occurred. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Keep the broader provisional event type. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Classification status and analyst notes. Unsupported geological causation enters the report. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 50: Focal mechanism

Fictional source situation. A moment-tensor display has two nodal planes. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The translation presents both as confirmed ruptured faults. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Describe possible fault planes rather than verified rupture geometry. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Moment tensor methodology and uncertainty. Alternative models become false facts. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 51: Strike dip rake

Fictional source situation. A mechanism lists strike, dip and rake in degrees. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. A table converts rake into geographic compass direction. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Keep each angle with its proper geometric definition. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Mechanism conventions and unit labels. Fault kinematics are misrepresented. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 52: Moment units

Fictional source situation. Seismic moment is reported in newton-metres. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The caption calls its number a magnitude directly. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Distinguish moment and derived moment-magnitude calculation. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Moment tensor properties and unit convention. A physical moment becomes a dimensionless scale. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 53: Intensity map

Fictional source situation. A product includes instrumental or reported shaking intensity by area. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The target describes all cells as earthquake magnitude. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Retain local intensity scale and spatial dependence. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Map legend and intensity method. Reader confuses source size with shaking effects. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 54: PGA versus PGV

Fictional source situation. A product lists peak ground acceleration and peak ground velocity. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The translation uses one phrase ‘peak strength’ for both columns. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Distinguish quantity and units in each column. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Product schema, axes and units. Engineering comparisons become unsound. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 55: ShakeMap uncertainty

Fictional source situation. A modeled map combines station observations and interpolation. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The translation calls every pixel a direct sensor reading. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Label estimates and mapping assumptions. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check ShakeMap technical description and product metadata. Spatial models are mistaken for measurements. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 56: Station coverage

Fictional source situation. An event solution uses four nearby stations and distant additional stations. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The target says four earthquakes were detected. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Read ‘used stations’ as an input count. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Origin quality metrics and station list. Coverage becomes occurrence frequency. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 57: RMS residual

Fictional source situation. An origin reports RMS travel-time residual 0.7 s. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The summary calls the earthquake’s duration 0.7 seconds. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Keep RMS as fit diagnostic tied to travel-time residuals. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Solution-quality definition. Physical event duration is invented. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 58: Azimuthal gap

Fictional source situation. The location report gives azimuthal gap 170 degrees. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The translation says a rupture gap of 170 degrees occurred. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Identify network coverage geometry, not rupture. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Origin-quality field definition. Location robustness is misrepresented. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 59: Station count

Fictional source situation. The field nst counts contributing stations under a provider convention. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The translation treats nst as a count of aftershocks. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Retain the provider-specific metric definition. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check USGS/producer schema and contribution notes. Seismic sequence size is fabricated. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 60: Update timestamp

Fictional source situation. A catalogue item was updated 30 minutes after origin time. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The text says the quake occurred twice. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Differentiate occurrence time from record update. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Event time and updated field. Chronology creates a phantom second event. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 61: Reviewed status

Fictional source situation. A record says reviewed, but uncertainty values remain. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The translation says location and magnitude are indisputable. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Retain review as a process status, not epistemic perfection. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Evaluation status and quality fields. Scientific uncertainty disappears. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 62: Data product version

Fictional source situation. A finite-fault solution is superseded by a later version. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The target cites an earlier graph as the current result. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Label version and processing date in the caption. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Product version and preferred flag. Reproducibility and interpretation diverge. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 63: StationXML output

Fictional source situation. An API response lists network and channel response information. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The translated help page calls the response raw waveform data. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Explain that this endpoint returns metadata. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check fdsnws-station service documentation. Users search in the wrong data product. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 64: Dataselect query

Fictional source situation. A request supplies net, sta, loc, cha and time window. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. A localized tutorial translates the parameter names. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Leave API keys and values in their documented syntax. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check FDSN service specification and test output. The example request stops working. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 65: Event query

Fictional source situation. A request for event parameters yields origin and magnitude records. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The guide promises every waveform for the event. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. State event parameters are not the complete waveform archive. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check fdsnws-event output documentation. Researchers expect unavailable samples. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 66: Availability query

Fictional source situation. A service reports the span of archived data for a channel. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The translation promises scientifically usable continuous recording. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Differentiate nominal availability from quality and continuity. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check FDSN availability response and gaps. Sampling omissions are hidden. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 67: miniSEED 3 version

Fictional source situation. An archive explicitly labels the record format miniSEED 3. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The target documentation describes it as identical byte-for-byte to SEED 2.4. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Retain version and relevant header differences. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check FDSN miniSEED 3 specification. A parser or conversion may misread records. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 68: miniSEED integrity

Fictional source situation. A record has a checksum/CRC-related integrity field under the chosen format. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The translation calls it instrument accuracy. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Describe integrity checks separately from measurement uncertainty. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Format documentation and file verification. File integrity becomes a false geophysical quality claim. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 69: Source Identifier migration

Fictional source situation. A legacy channel key is mapped to its FDSN Source Identifier form. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The translator removes a blank location separator to improve appearance. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Preserve the documented mapping including empty components. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check FDSN Source Identifier specification. The stable archival identity is damaged. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 70: SeedLink packet

Fictional source situation. A live waveform stream delivers a segment before analyst review. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The caption calls the packet a verified earthquake warning. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Describe streaming observation, not reviewed event status. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Protocol and processing-latency record. Public users may act on unreviewed content. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 71: Archive reprocessing

Fictional source situation. An archive replaces a corrupted segment with a corrected copy. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The translation calls both copies independent new observations. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Preserve version and remediation provenance. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Repository correction notice and hash history. Analytical samples are duplicated. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 72: Historical printed pick

Fictional source situation. A scanned bulletin uses a station-specific historical phase abbreviation. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. An editor silently normalizes it as a modern P phase. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Transcribe the historic symbol and annotate uncertain mapping. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Digitized scan and archival catalogue. The original observer’s claim is rewritten. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 73: Optical image of a seismogram

Fictional source situation. A chart image shows pen traces from a historic instrument. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The alt text states a calibrated velocity value not printed in source. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Describe visible features and limitations only. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Original scan scale and archive record. Accessibility text invents quantification. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 74: Restricted network

Fictional source situation. A station metadata entry is public but waveform access is restricted. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The translation calls the complete dataset open access. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. State access and reuse conditions separately. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Repository permissions and license. Researchers receive a false access promise. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 75: Dataset DOI

Fictional source situation. A seismic network archive has a DOI citation and named contributors. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The target handbook drops its citation in translation. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Preserve the DOI and required attribution. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check FDSN DOI recommendations and repository citation. Scholarly provenance disappears. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 76: Tsunami flag

Fictional source situation. An event product includes an agency-specific indicator related to tsunami products. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The translation interprets the raw flag as a confirmed wave at every coast. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Explain the exact field meaning without inventing hazard conclusions. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Provider product definitions and responsible warning authority. A technical flag becomes an unauthorized public warning. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 77: Map screenshot

Fictional source situation. A catalogue map was captured before the latest origin revision. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The translated caption gives current coordinates without updating the image. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. State the map’s time/version or refresh under controlled review. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Image timestamp and event product history. Text and image imply contradictory facts. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 78: Significance metric

Fictional source situation. An API supplies a composite significance score next to magnitude. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The translation labels the score as the event’s moment magnitude. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Keep significance and magnitude separately named. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check API field dictionary. Ranking values become scientifically impossible magnitudes. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 79: Aftershock label

Fictional source situation. A catalog editor calls an event an aftershock only in a later sequence analysis. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. A real-time translated headline uses the label before analysis. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Preserve classification timing and source attribution. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Sequence study and catalogue revision. A retrospective relationship is treated as established immediately. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Diagnostic 80: Cross-border data exchange

Fictional source situation. Two networks publish an event in different languages with different preferred solutions. This record must be understood as a claim made at a particular layer of the seismic evidence chain. Before translating it, identify the data object, its time or epoch, its measurement status, any controlled codes and the organization or process responsible for the value.

Plausible but incorrect interpretation. The localizer chooses one without citing the producing agency. A fluent sentence could hide this substitution because the substituted phrase looks familiar to an ordinary reader. The translator should ask whether the new wording changes who observed the fact, how the quantity was obtained, or whether the assertion is preliminary, derived, reviewed, unavailable or merely descriptive.

Defensible repair. Show which organization, time and model each value belongs to. Keep an immutable version of any machine-readable field and add a human-readable gloss only in a separate explanatory location. Preserve quantitative precision when justified by the source; never manufacture greater certainty, additional observations or an implied public instruction. The target should support the same record lookup and scientific inference as the source.

Reviewer proof and consequence. Check Agency IDs and publication records. Readers think one translated claim represents global scientific consensus. To close the issue, the reviewer should record the source reference, the rejected wording, the accepted wording and a simple falsification test: what observable mismatch would appear if the incorrect translation were used? This is stronger than marking the sentence ‘checked’, because it shows why the decision is correct and how a later translator can reproduce it.

Field workshops: twelve transfer challenges

Transfer workshop 1: From a press release to a catalogue

Assignment. Translate a fictional press summary ‘magnitude 5 nearby, later revised’ into a technical source ledger with event ID, magnitude type, origin time, agency, revision and uncertainty. The trainee begins by copying the source without editing its codes or units, then writes a target version for a named reader. The evaluator does not grade elegance until the evidence objects and their relations are preserved.

Acceptance test. The ledger must not invent data absent from the release; identify the fields requiring retrieval from the authoritative event service. Review the result in three passes: factual sameness, reproducibility and readability. A failed factual or reproducibility test overrides stylistic polish. Keep the rejected alternative and a short explanation of the changed claim so that the learning transfers to the next data type.

Transfer workshop 2: From legacy NSLC to FDSN SID

Assignment. Take a fictional network/station/empty-location/BHZ tuple and map it to a source identifier while preserving the empty location delimiter. The trainee begins by copying the source without editing its codes or units, then writes a target version for a named reader. The evaluator does not grade elegance until the evidence objects and their relations are preserved.

Acceptance test. Check that the inverse mapping, when supported, restores the exact original tuple and does not insert location 00. Review the result in three passes: factual sameness, reproducibility and readability. A failed factual or reproducibility test overrides stylistic polish. Keep the rejected alternative and a short explanation of the changed claim so that the learning transfers to the next data type.

Transfer workshop 3: From digital counts to physical meaning

Assignment. Explain to a non-specialist why a waveform axis in counts is not ground acceleration and why instrument response metadata matters. The trainee begins by copying the source without editing its codes or units, then writes a target version for a named reader. The evaluator does not grade elegance until the evidence objects and their relations are preserved.

Acceptance test. Produce a two-sentence readable explanation plus a technical reviewer note; do not compute units without a documented response. Review the result in three passes: factual sameness, reproducibility and readability. A failed factual or reproducibility test overrides stylistic polish. Keep the rejected alternative and a short explanation of the changed claim so that the learning transfers to the next data type.

Transfer workshop 4: From StationXML to prose

Assignment. Write an accessible channel description from station epochs, sensor orientation, sample rate and units. The trainee begins by copying the source without editing its codes or units, then writes a target version for a named reader. The evaluator does not grade elegance until the evidence objects and their relations are preserved.

Acceptance test. State which date range the description applies to and retain original channel codes. Review the result in three passes: factual sameness, reproducibility and readability. A failed factual or reproducibility test overrides stylistic polish. Keep the rejected alternative and a short explanation of the changed claim so that the learning transfers to the next data type.

Transfer workshop 5: From a revised event to two audiences

Assignment. Describe an automatic first solution and a later reviewed solution without implying that two earthquakes happened. The trainee begins by copying the source without editing its codes or units, then writes a target version for a named reader. The evaluator does not grade elegance until the evidence objects and their relations are preserved.

Acceptance test. Create a short public note and a detailed analyst change ledger, each with version times and provenance. Review the result in three passes: factual sameness, reproducibility and readability. A failed factual or reproducibility test overrides stylistic polish. Keep the rejected alternative and a short explanation of the changed claim so that the learning transfers to the next data type.

Transfer workshop 6: From a missing waveform to cautious reporting

Assignment. Explain a six-second missing interval in a report that would otherwise call the period seismically quiet. The trainee begins by copying the source without editing its codes or units, then writes a target version for a named reader. The evaluator does not grade elegance until the evidence objects and their relations are preserved.

Acceptance test. Distinguish missing recording from non-observation and propose a way to test additional channels. Review the result in three passes: factual sameness, reproducibility and readability. A failed factual or reproducibility test overrides stylistic polish. Keep the rejected alternative and a short explanation of the changed claim so that the learning transfers to the next data type.

Transfer workshop 7: From moment tensor to editorial caption

Assignment. Caption a focal-mechanism beachball diagram without asserting that either nodal plane is proven to be the ruptured fault. The trainee begins by copying the source without editing its codes or units, then writes a target version for a named reader. The evaluator does not grade elegance until the evidence objects and their relations are preserved.

Acceptance test. State model context and alternative interpretations at an audience-appropriate reading level. Review the result in three passes: factual sameness, reproducibility and readability. A failed factual or reproducibility test overrides stylistic polish. Keep the rejected alternative and a short explanation of the changed claim so that the learning transfers to the next data type.

Transfer workshop 8: From map to data table

Assignment. Cross-check a location map against signed latitude, longitude and depth without changing their reference frames. The trainee begins by copying the source without editing its codes or units, then writes a target version for a named reader. The evaluator does not grade elegance until the evidence objects and their relations are preserved.

Acceptance test. Test coordinate order, hemisphere signs and whether an old map version needs annotation. Review the result in three passes: factual sameness, reproducibility and readability. A failed factual or reproducibility test overrides stylistic polish. Keep the rejected alternative and a short explanation of the changed claim so that the learning transfers to the next data type.

Transfer workshop 9: From network comparison to provenance

Assignment. Combine two agency event reports without replacing different magnitude types or deleting contributing network IDs. The trainee begins by copying the source without editing its codes or units, then writes a target version for a named reader. The evaluator does not grade elegance until the evidence objects and their relations are preserved.

Acceptance test. A successful bilingual table shows each agency, method, value, status and timestamp. Review the result in three passes: factual sameness, reproducibility and readability. A failed factual or reproducibility test overrides stylistic polish. Keep the rejected alternative and a short explanation of the changed claim so that the learning transfers to the next data type.

Transfer workshop 10: From early signal to public message

Assignment. Explain what cannot be concluded from a live SeedLink packet before event processing and authorized public warning. The trainee begins by copying the source without editing its codes or units, then writes a target version for a named reader. The evaluator does not grade elegance until the evidence objects and their relations are preserved.

Acceptance test. Produce an educational text that is clear but contains no unsupported safety directive. Review the result in three passes: factual sameness, reproducibility and readability. A failed factual or reproducibility test overrides stylistic polish. Keep the rejected alternative and a short explanation of the changed claim so that the learning transfers to the next data type.

Transfer workshop 11: From archival scan to digital table

Assignment. Transcribe an old phase bulletin where column headings and abbreviations differ from modern usage. The trainee begins by copying the source without editing its codes or units, then writes a target version for a named reader. The evaluator does not grade elegance until the evidence objects and their relations are preserved.

Acceptance test. Retain original characters, flag unreadable values and place proposed modern mappings in separate annotated columns. Review the result in three passes: factual sameness, reproducibility and readability. A failed factual or reproducibility test overrides stylistic polish. Keep the rejected alternative and a short explanation of the changed claim so that the learning transfers to the next data type.

Transfer workshop 12: From translated article back to software

Assignment. Audit an article against its source event and dataselect query and reconstruct the exact original records. The trainee begins by copying the source without editing its codes or units, then writes a target version for a named reader. The evaluator does not grade elegance until the evidence objects and their relations are preserved.

Acceptance test. If another researcher cannot retrieve the same dataset or understand the same uncertainty, open a corrective translation issue. Review the result in three passes: factual sameness, reproducibility and readability. A failed factual or reproducibility test overrides stylistic polish. Keep the rejected alternative and a short explanation of the changed claim so that the learning transfers to the next data type.

Four-week practice and implementation plan

Week one: recognize data layers

Read the source descriptions of miniSEED, StationXML and QuakeML. Build a sheet distinguishing one physical event, three station channels, one station response epoch, a pick and a catalogue origin. Add a column for whether an item is observed, estimated or narrative. Use fictional records, and ensure every label points to a real concept. By the end of the week, demonstrate that no waveform sample has been confused with a magnitude estimate.

Deliverable: a verified concept map with source links, five contrastive examples, a protected-identifier column and a one-page explanation for non-specialists. Success is demonstrated by a reproducible decision and by the ability to explain why a tempting alternative translation was rejected. If a field cannot be verified, that uncertainty is an honest outcome rather than permission to fill the blank with intuition.

Week two: preserve structure and numbers

Work through source identifiers, empty location codes, timezones, sample rates, physical units, response-correction labels, spatial coordinates and uncertainty. Create automated text comparisons for protected strings where possible. Translate explanatory captions into a second language only after documenting the data values. Then deliberately insert mistakes such as a lost negative sign, a renamed XML key and a timestamp without UTC. Your checker should catch each one.

Deliverable: a bilingual field ledger, a successful round-trip identifier test, a numerical QA sheet and a list of unresolved questions for a domain reviewer. Success is demonstrated by a reproducible decision and by the ability to explain why a tempting alternative translation was rejected. If a field cannot be verified, that uncertainty is an honest outcome rather than permission to fill the blank with intuition.

Week three: revise event stories without changing evidence

Construct a fictional catalogue revision chain with automatic and reviewed origins, two magnitude types, an alternative association and changing publication timestamps. Write both a public summary and a research appendix. Keep a change log of where your target sentence became more or less specific. Invite a reviewer to challenge each sentence by asking what source field actually licenses it.

Deliverable: two audience versions that do not contradict the source, a reviewer sign-off log and a differential table of preliminary and later solutions. Success is demonstrated by a reproducible decision and by the ability to explain why a tempting alternative translation was rejected. If a field cannot be verified, that uncertainty is an honest outcome rather than permission to fill the blank with intuition.

Week four: test transfer and release control

Exchange sample material with another translator and ask them to reproduce the same station query and catalogue interpretation using your target text alone. Run failure-mode challenges: duplicated samples, gap called calm, preliminary called final, missing data called absence of events and magnitude type omitted. Finish with a release decision noting any limitation that remains. Test whether a new future revision can be incorporated without rewriting the historical evidence.

Deliverable: an audited publish-ready package containing a source snapshot, structured vocabulary, QA report, reviewer decision, correction policy and brief reader-facing explanation. Success is demonstrated by a reproducible decision and by the ability to explain why a tempting alternative translation was rejected. If a field cannot be verified, that uncertainty is an honest outcome rather than permission to fill the blank with intuition.

Glossary: terms that must remain attached to the right concept

event: interpreted object describing a candidate or accepted physical seismic occurrence. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

waveform: time-indexed sample sequence from an observing channel. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

station: identified observing site within a seismic network and epoch. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

network: registered namespace and operator grouping for station data. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

channel: identified stream with measurement and component semantics. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

location code: channel grouping component that may be empty and remain significant. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

source identifier: FDSN URI-like composite identifier for a data source. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

epoch: time interval during which a defined instrument or metadata configuration applies. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

StationXML: XML representation of network, station, channel and response metadata. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

miniSEED: binary time-series interchange and archival format. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

miniSEED 3: version of miniSEED with defined newer headers and record features. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

SEED 2.4: earlier seismological data exchange convention and associated codes. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

QuakeML: structured XML representation of seismological event objects. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

publicID: persistent object identifier used for links within event representations. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

preferredOriginID: link designating the preferred origin within an event object. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

origin: estimated event location and time under a fitted solution. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

hypocentre: subsurface source location including depth. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

epicentre: projection of source location to Earth’s surface. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

depth: vertical origin parameter requiring a stated reference and method. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

fixed depth: depth constrained by the solution procedure rather than freely estimated. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

magnitude: quantitative estimate of earthquake source size under a specified type. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

ML: local magnitude type. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

mb: body-wave magnitude type. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

Ms: surface-wave magnitude type. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

Mw: moment magnitude family label. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

seismic moment: physical source parameter distinct from a dimensionless magnitude. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

moment tensor: model of seismic source motion and orientation. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

focal mechanism: description of source radiation and possible fault geometry. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

nodal plane: one of possible plane orientations in a double-couple interpretation. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

strike: horizontal orientation of a plane under a geometric convention. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

dip: inclination of a plane under a geometric convention. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

rake: direction of movement on a plane under an angular convention. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

pick: estimated arrival time and related metadata for a seismic phase. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

P phase: compressional-wave arrival designation. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

S phase: shear-wave arrival designation. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

phaseHint: suggested seismic phase designation on a pick. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

arrival residual: difference between observed and modeled arrival timing. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

station distance: distance or angular separation used in event analysis. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

azimuth: direction angle with stated frame and origin. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

RMS residual: aggregate fit diagnostic for arrival times. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

azimuthal gap: network observation geometry measure for an origin. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

horizontal uncertainty: statistical uncertainty in horizontal event location. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

depth uncertainty: statistical uncertainty in estimated origin depth. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

evaluationMode: indicator of automatic or manual evaluation process where supplied. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

evaluationStatus: assessment or review state, distinct from update time. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

preferred magnitude: provider-designated magnitude estimate for display or analysis. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

origin time: estimated time at seismic source, not waveform start time. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

pick time: estimated signal arrival time at a station. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

updated: record or product modification timestamp. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

UTC: Coordinated Universal Time used to label event chronology. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

time quality: information about the reliability of a recording clock. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

sampling rate: number of samples per unit time. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

sample period: time interval between nominal samples. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

digital counts: unconverted digitizer values. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

instrument response: frequency-dependent transformation between physical input and recorded output. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

sensitivity: defined input/output relationship for an instrument or stage. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

gain: signal scaling factor under documented instrumentation context. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

response removal: processing step that estimates physical motion from recorded data. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

counts to velocity: calibration-dependent conversion, not an automatic relabeling. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

displacement: physical positional change with units of length. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

velocity: rate of displacement with units of length per time. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

acceleration: rate of velocity change with units of length per time squared. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

clipping: signal saturation limiting the recorded amplitude. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

data gap: interval without available valid samples. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

overlap: time interval represented more than once in archive content. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

noise: non-target or unwanted signal affecting observation. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

focal depth: another name for earthquake origin depth in some contexts. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

intensity: place-specific shaking or effect measure on an identified scale. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

MMI: Modified Mercalli Intensity scale for shaking effects. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

PGA: peak ground acceleration. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

PGV: peak ground velocity. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

ShakeMap: mapped estimate of ground shaking under a documented product model. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

fdsnws-station: web service for station metadata. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

fdsnws-dataselect: web service for waveform time series. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

fdsnws-event: web service for seismic event parameters. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

fdsnws-availability: web service for waveform data-availability metadata. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

SeedLink: streaming exchange protocol for seismological time-series data. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

ComCat: USGS-operated comprehensive event catalogue and associated products. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

ISC Bulletin: International Seismological Centre event and phase data resource. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

archive provenance: recorded origin and transformations of retained scientific data. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

DOI: persistent digital identifier useful for citing a dataset. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

data licence: conditions governing allowed access or reuse. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

release gate: documented pass/fail check before public or technical distribution. Its translation should be checked against the specific producer’s data dictionary and the sentence’s evidence role; terminology alone never overrides the original field or method.

Frequently asked questions

Can an earthquake catalogue be translated like ordinary prose?

Only the explanatory human language can be freely rewritten. Event identifiers, quantities, units, field keys, timestamps and statuses need controlled treatment. In a scientific handover, successful translation should still let a user locate and interpret the same event parameters. For a high-consequence use, the reviewer should document a specific source artifact and what the target reader is permitted to infer. When evidence is incomplete, answer conditionally rather than guessing.

What is the difference between waveform translation and event translation?

Waveform records describe time-series samples with channel and instrument context. Event records summarize interpreted locations, times, magnitudes and associated picks. They are related but neither is a synonym for the other. For a high-consequence use, the reviewer should document a specific source artifact and what the target reader is permitted to infer. When evidence is incomplete, answer conditionally rather than guessing.

What is FDSN?

The International Federation of Digital Seismograph Networks publishes data exchange specifications and associated identifier conventions used across seismological networks. The exact specification and version should be cited when a workflow depends on them. For a high-consequence use, the reviewer should document a specific source artifact and what the target reader is permitted to infer. When evidence is incomplete, answer conditionally rather than guessing.

What is StationXML?

A metadata schema for networks, stations, channels and responses. Its descriptive text may be translated in documentation; its structural keys and protected data values are not free text. For a high-consequence use, the reviewer should document a specific source artifact and what the target reader is permitted to infer. When evidence is incomplete, answer conditionally rather than guessing.

What is miniSEED 3?

A standardized time-series data format used for seismological exchange. It adds defined format capabilities relative to earlier miniSEED, so a translation must retain the exact version rather than treating all files as the same bytes. For a high-consequence use, the reviewer should document a specific source artifact and what the target reader is permitted to infer. When evidence is incomplete, answer conditionally rather than guessing.

Is QuakeML a human-readable article?

No. It is a structured event data representation with identifiers and object links. A translator may explain its content, but changing identifiers and schema tags can break machine interpretation. For a high-consequence use, the reviewer should document a specific source artifact and what the target reader is permitted to infer. When evidence is incomplete, answer conditionally rather than guessing.

Does an updated earthquake report mean it was reviewed?

No. The updated time tells you that a record or product changed. Human review, evaluation mode and preferred solution status have separate meanings and should be verified separately. For a high-consequence use, the reviewer should document a specific source artifact and what the target reader is permitted to infer. When evidence is incomplete, answer conditionally rather than guessing.

Are magnitude and intensity the same?

No. Magnitude characterizes estimated earthquake source size using a specified method. Intensity describes shaking experience or effects at a place by an intensity scale or model. For a high-consequence use, the reviewer should document a specific source artifact and what the target reader is permitted to infer. When evidence is incomplete, answer conditionally rather than guessing.

Can ML automatically become Mw?

No. Magnitude types have different methods and applicable conditions. A source ML value should keep its type unless a separate scientifically justified conversion is made and labeled. For a high-consequence use, the reviewer should document a specific source artifact and what the target reader is permitted to infer. When evidence is incomplete, answer conditionally rather than guessing.

Does a data gap mean no shaking occurred?

No. It means the relevant observations were unavailable or invalid for that interval. An absence of recording is not an observation of absence. For a high-consequence use, the reviewer should document a specific source artifact and what the target reader is permitted to infer. When evidence is incomplete, answer conditionally rather than guessing.

Is raw digital count a unit of ground acceleration?

No. Counts are instrument output values. Physical-unit estimates require correct instrument response information and documented processing. For a high-consequence use, the reviewer should document a specific source artifact and what the target reader is permitted to infer. When evidence is incomplete, answer conditionally rather than guessing.

Why keep an empty location code?

Because it may be a valid part of channel identity. Replacing an empty code with 00 changes the query and may select another stream. For a high-consequence use, the reviewer should document a specific source artifact and what the target reader is permitted to infer. When evidence is incomplete, answer conditionally rather than guessing.

What is a preferred origin?

An origin chosen for use or display by the data provider from potentially several alternatives. It is not a guarantee that all other solutions are false or that uncertainty has vanished. For a high-consequence use, the reviewer should document a specific source artifact and what the target reader is permitted to infer. When evidence is incomplete, answer conditionally rather than guessing.

Can an event have multiple catalogue IDs?

Yes. Networks and aggregators may assign different identifiers to one occurrence. Cross references and provenance must be examined before deduplicating or counting events. For a high-consequence use, the reviewer should document a specific source artifact and what the target reader is permitted to infer. When evidence is incomplete, answer conditionally rather than guessing.

Should station names be translated?

Descriptions and common geographic labels can be explained where appropriate. Machine station codes should remain intact, and historic aliases should not replace registered identifiers. For a high-consequence use, the reviewer should document a specific source artifact and what the target reader is permitted to infer. When evidence is incomplete, answer conditionally rather than guessing.

How should local times be displayed?

Keep a canonical, unambiguous source timestamp, normally UTC in seismic exchange, and label any converted display as derived with its timezone. Do not change event ordering. For a high-consequence use, the reviewer should document a specific source artifact and what the target reader is permitted to infer. When evidence is incomplete, answer conditionally rather than guessing.

How can a translator verify a channel?

Use the network, station, location, channel, valid epoch and instrument-response metadata. The same station label can cover multiple historical configurations. For a high-consequence use, the reviewer should document a specific source artifact and what the target reader is permitted to infer. When evidence is incomplete, answer conditionally rather than guessing.

What is a phase pick?

An estimated arrival-time record for a seismic phase, produced manually or automatically. Its existence alone is not evidence that an earthquake solution has been finally validated. For a high-consequence use, the reviewer should document a specific source artifact and what the target reader is permitted to infer. When evidence is incomplete, answer conditionally rather than guessing.

Can AI translate seismology records safely?

AI can help draft explanations and detect inconsistent labels, but protected values and scientific interpretations require source-controlled checks. A domain reviewer should test whether AI introduced confidence or measurements absent from the record. For a high-consequence use, the reviewer should document a specific source artifact and what the target reader is permitted to infer. When evidence is incomplete, answer conditionally rather than guessing.

What is the best quality metric?

For this task, a translation must preserve identity, numeric evidence, provenance and the permitted inference. A general fluency score alone is insufficient; error-based evaluation and explicit release criteria are needed. For a high-consequence use, the reviewer should document a specific source artifact and what the target reader is permitted to infer. When evidence is incomplete, answer conditionally rather than guessing.

Why cite both a data provider and a standard?

A specification defines general structure, while a provider’s product and method documentation determine how its actual fields are populated. Both can matter to the reader’s interpretation. For a high-consequence use, the reviewer should document a specific source artifact and what the target reader is permitted to infer. When evidence is incomplete, answer conditionally rather than guessing.

Are historic printed bulletins obsolete?

They remain important evidence, often with distinct instruments and notation. The translator should preserve the original transcription and label editorial mappings rather than silently rewriting history. For a high-consequence use, the reviewer should document a specific source artifact and what the target reader is permitted to infer. When evidence is incomplete, answer conditionally rather than guessing.

Should a translated example be treated as a live safety warning?

No. Fictional exercises teach record interpretation, not urgent instructions. Use real alerts only from competent authorities and retain their exact time, geography and status. For a high-consequence use, the reviewer should document a specific source artifact and what the target reader is permitted to infer. When evidence is incomplete, answer conditionally rather than guessing.

How do I correct a translated seismic article after a catalogue revision?

Record which source version was previously used, compare changed fields, update the relevant target claims, and keep an audit trail. Do not erase historical values without explaining their chronology. For a high-consequence use, the reviewer should document a specific source artifact and what the target reader is permitted to infer. When evidence is incomplete, answer conditionally rather than guessing.

Authoritative sources, related study and next routes

The technical source trail below is designed to let a reader verify terminology at its actual origin. Links point to standards bodies and scientific data providers, not copied passages. The text paraphrases key concepts and gives original fictional examples. Source specifications have their own release dates; a page being crawled or checked in 2026 does not make an older version a 2026 edition.

FDSN Web Services specification and versions. Check the issuer’s currently applicable version and the data producer’s own method before relying on any operational interpretation.

FDSN Source Identifiers specification. Check the issuer’s currently applicable version and the data producer’s own method before relying on any operational interpretation.

FDSN channel codes. Check the issuer’s currently applicable version and the data producer’s own method before relying on any operational interpretation.

FDSN StationXML schema. Check the issuer’s currently applicable version and the data producer’s own method before relying on any operational interpretation.

FDSN StationXML version 1.2 background. Check the issuer’s currently applicable version and the data producer’s own method before relying on any operational interpretation.

FDSN miniSEED 3 specification. Check the issuer’s currently applicable version and the data producer’s own method before relying on any operational interpretation.

FDSN miniSEED 3 version background. Check the issuer’s currently applicable version and the data producer’s own method before relying on any operational interpretation.

FDSN publications and data citations. Check the issuer’s currently applicable version and the data producer’s own method before relying on any operational interpretation.

International Seismological Centre Bulletin and data resources. Check the issuer’s currently applicable version and the data producer’s own method before relying on any operational interpretation.

USGS ANSS Comprehensive Earthquake Catalog (ComCat). Check the issuer’s currently applicable version and the data producer’s own method before relying on any operational interpretation.

USGS ComCat documentation. Check the issuer’s currently applicable version and the data producer’s own method before relying on any operational interpretation.

USGS earthquake magnitude types. Check the issuer’s currently applicable version and the data producer’s own method before relying on any operational interpretation.

USGS GeoJSON earthquake summary fields. Check the issuer’s currently applicable version and the data producer’s own method before relying on any operational interpretation.

ISO 5060:2024 translation output evaluation guidance. Check the issuer’s currently applicable version and the data producer’s own method before relying on any operational interpretation.

Recommended route. Return to Master Art of Translation | The Complete System for Moving Meaning Between Languages for general translation architecture and the distinction between a source text, a technical concept and a release decision. The protected Vocabulary Learning Hub is relevant when a learner needs to practice word families such as estimate, observation, uncertainty, signal, measurement, event, evidence and inference; this chapter does not duplicate its learning lists. The protected How English Works | Translation and Meaning Across Languages is relevant for sentence-level explanation, qualification and reader comprehension; it remains the owner of that broader English-language lesson.

For the narrower task of writing reader-facing earthquake bulletins, magnitude-scale explanations or ShakeMap captions, follow the existing focused earthquake bulletin translation guide. For archival seismic metadata, provenance, identifiers and reproducibility, use this chapter. These two routes answer different questions and should not be treated as interchangeable search results.

Final transfer principle. A translation is defensible when an independent reader can still identify the same event, retrieve the same records, understand the same limitations and make only those claims the data actually support. That is how language moves scientific meaning without moving the evidence underneath it.