Audit actual USGS phase records before interpreting the interior

1. Define the evidence and frame

This destination uses actual public USGS phase-data records for the 6 February 2023 earthquake identified as us6000jllz, not invented seismic readings. The versioned QuakeML product supplies a preferred origin at 01:17:34.342 UTC, station picks, phase labels, angular distances and model residuals. The downloadable extract retains six P picks whose interpretation agency is us, selected near six distance targets without selecting on residual size.

2. Connect mechanism and model

Derived travel time is pick UTC minus origin UTC. A residual is observed travel time minus a model-predicted travel time; it is not the pick uncertainty and does not directly measure a local rock-speed error. At station KR.JNKS, travel is 369.378 s and catalog residual is +0.12 s, implying predicted travel 369.258 s for this version.

The sensitivity control moves the assumed origin later by b seconds while retaining the pick and model-predicted travel time. Both derived travel and residual decrease by b. This is an explicit hypothetical calculation, not a relocated earthquake, updated USGS record or claim that origin adjustment is uniquely justified. Actual relocation would reconsider location, depth, travel model and many observations jointly.

The records represent one event at six stations, not six independent earthquakes. Surface angular separation is not the length of the refracted interior ray. Do not divide surface distance by travel time and report a measured interior speed. Catalog picks are interpreted data, not raw waveforms or known truth; retain selection, agency, version, units, source hash and limitations before inference.

3. Reconstruct the worked packet

For KR.JNKS, 01:23:43.720−01:17:34.342=369.378 s. Subtracting residual 0.12 gives predicted 369.258 s. A hypothetical origin shift +0.5 s changes travel to 368.878 s and fixed-model residual to −0.38 s; raw data stay unchanged.

Actual catalog evidence: geo-usgs-phase-subset@1

2023-02-06 Pazarcik earthquake, Kahramanmaras earthquake sequence. Original preferred-origin UTC: 2023-02-06T01:17:34.342Z. The derived travel-time column is calculated; the other displayed fields retain the source values.

Six selected P arrivals; all pick interpretation agencies are USGS (us)

Network.station	Pick UTC	Distance (degrees)	Travel (s)	Catalog residual (s)
KR.JNKS	2023-02-06T01:23:43.72Z	30.0247	369.378	0.12
GE.KIBK	2023-02-06T01:25:04.91Z	39.3953	450.568	0.48
QM.MBUZ	2023-02-06T01:26:31.75Z	50.3615	537.408	0.14
IC.BJT	2023-02-06T01:27:40.98Z	59.9224	606.638	0.44
MY.KOM	2023-02-06T01:28:49.75Z	70.5736	675.408	-0.22
IM.BCAR	2023-02-06T01:29:44.55Z	80.0451	730.208	0.75

Download the uniform-schema CSV · Download versioned provenance · Original USGS QuakeML product

Selection: Preferred origin; phase P; pick creation agency us; angular distance 30–90 degrees. Select nearest distance to each target 30,40,50,60,70,80 degrees; ties lexical pick ID. No residual-based exclusion. Six distinct picks.

Transformation: Original identifiers, station/channel, UTC pick, angular distance and residual retained. Derived travel seconds=pick UTC minus preferred-origin UTC, rounded to 0.001 s. All other selected numeric values unchanged.

Reuse: USGS-produced phase-pick/arrival interpretation selected by agency us; USGS data public domain, credit U.S. Geological Survey and contributing station networks. No USGS logo or third-party graphics included. USGS public-domain and third-party exceptions policy. No endorsement is implied.

Source SHA-256: b496d5184be2d1b41bea320b007ac29bd196c8beafbae577f9076765319560ab. Retrieved 2026-10-30; source product update identifier 1681594773040.

Catalog picks and model residuals, not raw waveform measurements or ground-truth travel times. Source location/origin and phase identification are interpreted; this subset is not globally representative and has one common earthquake.

Retained actual-data audit

Before public feedback, save the original source URL/version/hash, agency and reuse terms in the course evidence workspace. Reproduce one UTC subtraction and one residual-to-prediction conversion. Explain the distance-based selection rule, common-event dependence and at least two possible causes of residuals. Preserve raw rows separately from a hypothetical origin shift. A new independent analysis must use a genuinely unfamiliar authorized packet, not rename this public example.

4. Predict and investigate

Predict first: Would shifting the assumed origin later increase or decrease all fixed-model residuals?

Download the CSV and provenance, verify one UTC subtraction and compare all six residuals. Preserve the original rows and source license audit. Change only the hypothetical origin offset, then explain why this cannot by itself identify a revised origin or a unique Earth velocity model.

5. Test a changed case

Apply the evidence to a changed case

For IM.BCAR, original travel 730.208 s and residual +0.75 s, apply a hypothetical +1 s origin shift. What changes and what remains a source observation?

Retain the evidence frame, calculations and assumptions before opening public feedback.

Compare a worked response

Predicted travel remains 729.458 s; sensitivity travel is 729.208 s and residual −0.25 s. Original pick, catalog residual, angular distance and version remain unchanged in the source packet.

6. Retain a defensible evidence notebook

Preserve first predictions, full units, geometry, source provenance, assumptions and rejected alternatives. Synthetic models are not field measurements. The explicitly labeled USGS packet is actual catalog evidence, but working through a public example is not independent practical competence or an operational hazard assessment.

Mechanism and data reference, checked 30 October 2026. Original teaching prose, diagrams and exercises; actual USGS rows retain their agency and provenance.