The earthquake near Arthur’s Pass on Tuesday night has a useful lesson for anyone following a fast-moving event: identify the record before comparing the numbers. An early screenshot and a later official page can describe the same earthquake differently. Reposting the older value without its time and review status removes the very information needed to understand that difference.

At the research check for this article on 18 November 2026, GeoNet’s event page displayed magnitude 5.7, a depth of 5 km and a location 15 km south-east of Arthur’s Pass. It placed the earthquake at 11:30 pm on 17 November in New Zealand local time, with Strong shaking and Best quality. The event identifier is 2026p866048. These are a dated record check, not a promise that a scientific catalogue will never be revised. GeoNet event overview.

This guide explains how to keep those fields together, compare a superseded estimate with a reviewed one, and avoid turning a search trend into an unsupported fault or hazard claim. It is not a structural inspection, a forecast of another earthquake or a live emergency bulletin.

Start with one event identifier

A place name is convenient for conversation but a poor substitute for a catalogue identity. Two readers might search for the nearest familiar town, a region where they felt shaking, or the place mentioned in a headline. Their descriptions can sound different even when they are discussing one record. Conversely, several earthquakes can appear in a regional search without being the same event.

For this story, keep 2026p866048 beside the link. Copying that identifier into a research note provides a much firmer reference than writing “last night’s quake.” Tomorrow, last night refers to a different date. A forwarded message might be opened in a different time zone. A search result might place a smaller later event above the one that originally prompted the query.

A useful personal note has four separate entries: the event identifier, the origin time, the page address and the time you checked it. Add your own question below them. Perhaps you want to understand an apparent magnitude change; perhaps you want to find an official preparedness page. Keeping the question visible prevents unrelated search results from silently changing the subject.

For an international reader, include the time-zone label explicitly. The direct record uses 10:30:48 UTC on 17 November, while the public page presents the late-evening local time. Those are two presentations of the event time, not a second earthquake. Keeping UTC beside a locally formatted time is particularly helpful when relatives overseas compare messages.

The reviewed value and the earlier estimate

GeoNet’s history page makes the revision sequence visible. Its table includes an automatic magnitude 5.4 estimate at 10:34:35 UTC on 17 November, with depth 14.4 km. A reviewed Best entry at 10:35:20 UTC lists magnitude 5.7 and depth 5.0 km. Another Best entry at 10:42:42 UTC retains those magnitude and depth values while refining the coordinates. GeoNet describes Best entries as reviewed by a duty person. Event calculation history.

These rows are successive descriptions of one event. They are not evidence that the ground first produced one earthquake and then became a larger earthquake several minutes later. Notice that the history table labels its timestamp as a modification time. That is a different field from the event’s origin time.

During research, an older indexed technical result still showed an automatic 5.3 estimate. Opening the current overview and checking GeoNet’s direct event data resolved which value to use here. The direct data retrieval at 06:42 UTC on 18 November gave magnitude approximately 5.66, displayed by the overview to one decimal place as 5.7, with depth 5 km and quality best. GeoNet event data.

The practical correction is straightforward: retain the old screenshot only as evidence of what was displayed earlier, and label it accordingly. Use the reviewed record for a current description. Do not combine the later magnitude with an earlier depth simply because each number appears somewhere in the search results.

Why a changed estimate is not a contradiction to hide

GeoNet explains that initial event information is generated automatically and that magnitude and location can change as additional information arrives or an analyst reviews it. It also notes that different agencies may use different stations, models and methods. That is a reason to compare like with like, rather than choosing whichever number makes the strongest headline. GeoNet earthquake FAQ.

A useful comparison asks whether two figures came from the same agency, the same event and the same stage of review. If those conditions differ, name the difference before drawing a conclusion. It is possible to be transparent without burying readers in technical detail: “This earlier automatic estimate has been superseded in the current GeoNet record” says more than quietly replacing a number.

For a community newsletter, a correction note can be brief. State which field changed, which record now supports it, and when the new check occurred. Keep the original publication time separate from the correction time. Otherwise a newly edited story can appear to have contained the revised information from the beginning, obscuring how its account developed.

Keep magnitude, shaking and quality in different boxes

GeoNet distinguishes magnitude, which describes earthquake size, from intensity, which describes effects at ground level. Its simplified intensity scale associates MMI 6 with Strong shaking. The two labels answer different questions; one should not be substituted for the other. GeoNet intensity explanation.

The Strong label on this event’s overview is not a statement that every reader experienced identical shaking. Nor does the quality label Best mean that a particular building has been inspected or declared safe. Quality describes confidence in the event details. A building assessment would be a different kind of evidence, produced for a different purpose.

Imagine a reader making three columns headed “event size,” “shaking description” and “review status.” Putting each label in its own column exposes mistakes that a sentence can conceal. “Best shaking” would immediately look misplaced. So would using a magnitude figure as an inspection result for an individual home. The point of the exercise is to preserve the meaning of each field, not to create a new unofficial hazard scale.

A search for fault does not identify a fault

Google Trends New Zealand showed “fault” in the first displayed row when checked at 06:41 UTC on 18 November, using the past 24 hours and search-volume sorting. It displayed a 20K+ search band, a 1,000% increase and Active status. Related terms in that row included GeoNet and New Zealand earthquake searches. Those observations explain the editorial interest in this topic; they do not establish what every searcher meant or which geological structure moved. Google Trends New Zealand.

The distinction matters because a search phrase can tempt a writer to promise a more specific answer than the evidence supports. This article does not assign the event to a named fault. A location marker near a familiar feature is not, by itself, an expert attribution. Neither is an annotated social-media image whose original source cannot be checked.

If you are comparing a fault claim with the event overview, write the claim as a question first: “Which primary scientific statement identifies that structure?” A place name in a headline does not answer it. If the statement is unavailable, leave the attribution open rather than filling the gap with geographical guesswork.

A fictional example of correcting a shared message

Consider a fictional example. Maia helps maintain a neighbourhood information group. Someone posts an old screenshot and writes that officials “changed the earthquake.” Another person replies with a newer number, but leaves out the event link. The discussion now contains two claims and no shared reference.

Maia first checks that the screenshots concern the same event identifier. She then reads the reviewed entry and separates the origin time from the calculation update. Her replacement message gives the current source link and explains that the earlier image captured an automatic estimate. She does not accuse the first person of deliberately misleading anyone; the visible problem is missing context.

Finally, she removes a sentence in her own draft suggesting that the revision proves what will happen next. Nothing in the comparison supports that prediction. The corrected post becomes less dramatic but more useful: readers can reproduce the check, see what changed, and decide which official information they need next. This is an illustration of an editing method, not a report of a real group or person.

Do not merge separate notices into one story

When several hazard terms appear in search results on the same day, chronological proximity can look like a causal connection. A screenshot about one location may circulate beside a notice concerning another coast or another incident. A responsible summary keeps the event identifier and the issuing authority attached to each item instead of presenting a bundle of loosely related posts as one confirmed sequence.

For emergency instructions, consult the issuing authority’s current notice and the relevant local emergency information. The National Emergency Management Agency provides an emergency events starting point. This article does not certify the current status of any separate coastal advisory or announce an all-clear. A general article saved on a phone cannot substitute for a later official update.

Before sharing a notice, check its issue time, geographic scope, action requested and whether a later notice replaces it. Those four details are especially important when someone crops an image to fit a messaging app. If the crop removes them, share the original link instead. Avoid adding a cause that the issuing notice does not itself establish.

Keep immediate safety advice separate from record checking

Get Ready advises Drop, Cover and Hold during earthquake shaking, protecting the head and neck and using nearby shelter where possible. Its guidance says not to run outside during shaking. For people near the coast, shaking that lasts longer than a minute or makes standing difficult is a natural warning: once shaking stops, move promptly to high ground or inland beyond tsunami evacuation zones. Do not delay that response to check a magnitude on a website. Official earthquake guidance.

The rest of this article concerns interpreting information when it is safe to do so. Reading a history table is a useful later task, but it is not an action to prioritise while taking shelter. An accurate understanding of a record should support preparedness, not distract from the official protective steps that apply at the moment.

Make a short, reusable verification note

For a classroom discussion, family message or community update, prepare a compact note with headings rather than an unstructured collection of screenshots. Under “confirmed,” place only details that your named source supports. Under “superseded,” keep the earlier value and its context if explaining the change is necessary. Under “unresolved,” put questions such as a fault attribution that you have not verified.

This format makes revision easier. A later update can move a question from unresolved to confirmed, with a source and time attached. It also helps another person see the limits of your work. They do not have to guess whether silence about damage means there was none, or merely that your note did not investigate it. In this guide, no independently verified damage total is offered.

For a broader method of connecting a claim to its evidence, see our guide to turning a primary source into evidence you can explain. The useful habit is the same across subjects: preserve what the source actually supports and identify what remains an interpretation.

Conclusion: share a traceable update

The most useful account of the Arthur’s Pass earthquake gives readers a route back to the reviewed record. Keep the identifier, distinguish event time from update time, and explain why an automatic estimate may have been replaced. Treat search interest as interest, not geological evidence. Keep emergency instructions tied to official current guidance, and avoid unsupported predictions or fault names.

Sources consulted 18 November 2026. Event data were retrieved at 06:42 UTC; the Google Trends snapshot was retrieved at 06:41 UTC. These are research timestamps, not continuous monitoring. The illustration is an original conceptual artwork, not a photograph of this event, an official map or a scientific fault diagram.

Source / image note: GeoNet event overview, calculation history, direct event data, FAQ and intensity guidance; Get Ready earthquake guidance and NEMA, consulted 18 November 2026. Google Trends New Zealand retrieved 06:41:41 UTC, past 24 hours, search volume descending.