Why Earthquake Magnitudes Change After the First Report
An earthquake posts as M6.9, then becomes M7.1 an hour later. Here is why USGS magnitude, depth, and location estimates get revised, how much change is normal, and which number to trust.
Quick Answer: Why Do Earthquake Magnitudes Change?
The first magnitude you see is a computer estimate built from the nearest reporting stations, using only the first seconds of waveform. How quickly it appears depends on where the earthquake happened: inside a dense network like California's, USGS processes and posts an event in about 2.5 minutes on average, while for earthquakes elsewhere in the world it aims to release an initial estimate within about 20 minutes. From there, waves reach hundreds more stations, a human analyst reviews the solution, and better methods become available. The number moves because the evidence improves.
| What Changes | Why It Changes | Typical Shift |
|---|---|---|
| Magnitude | More stations report; better-suited magnitude scales become usable | 0.1 to 0.3 is routine; 0.5 or more for very large events |
| Depth | Depth is the hardest parameter to resolve from sparse early data | Can move tens of kilometers |
| Location | Additional arrival times triangulate the source more tightly | A few kilometers, occasionally more |
| Status | An automatic solution gets reviewed by a seismologist | "automatic" becomes "reviewed" |
A revision is not a correction of an error. It is the normal, expected refinement of a measurement that was published while it was still incomplete, because getting a rough number out quickly is more useful than a perfect number hours later.
The First Number Comes From Very Few Stations
When a fault ruptures, seismic waves spread outward at a few kilometers per second. Automatic systems at agencies like the USGS are designed to publish something fast, so they work with whatever has arrived: often the closest few stations, a handful of seconds of waveform, and an automatic pick of when the P-wave arrived.
That is enough to produce a location and a magnitude. It is not enough to produce a confident one. Early solutions can be skewed by:
- Station geometry. If every reporting station sits on one side of the epicenter, the triangulation is lopsided and both location and depth suffer.
- Bad automatic picks. Algorithms occasionally mistake noise, a passing truck, or the tail of a previous earthquake for a wave arrival.
- Overlapping events. In an active aftershock sequence, two earthquakes seconds apart can be merged into one oversized solution, or one event split into two.
As waves keep travelling, the station count climbs from a handful into the hundreds. Each new arrival time constrains the answer further.
Large Earthquakes Are Still Happening When the First Estimate Is Published
This is the single most important reason big magnitudes get revised upward, and it surprises most people.
An earthquake is not an instantaneous point event. It is a rupture tearing along a fault at roughly two kilometers per second, and the bigger the earthquake, the longer that takes:
| Magnitude | Approximate Rupture Length | Roughly How Long the Fault Keeps Breaking |
|---|---|---|
| M5 | A few kilometers | Under a second |
| M6 | ~10 km | A few seconds |
| M7 | ~50 km | 10 to 20 seconds |
| M8 | ~200 km | Around a minute |
| M9 | 500 to 1,000+ km | Several minutes |
An automatic system that computes a magnitude 60 seconds after the first waves arrive is, for a magnitude 9, measuring an earthquake that is still in progress. It cannot know how much more fault is about to break. So it reports the earthquake that has happened so far, and that number is too small.
This is why the world's largest earthquakes often appear first in the high M7s or low M8s, climb through several revisions, and settle a magnitude unit or more higher. USGS first posted the 2011 Tohoku earthquake at M7.9. It was raised to M8.8, then M8.9, then M9.0 within days, and it now stands at M9.1.
Some Magnitude Scales Physically Cannot Measure Big Earthquakes
The second reason large events get revised upward is saturation. Different magnitude scales measure different parts of the seismic wave, and most of them hit a ceiling.
| Scale | USGS Label | What It Measures | Saturates Around |
|---|---|---|---|
| Local magnitude | ml |
Peak amplitude on a local instrument | M6.5 to M7 |
| Body-wave magnitude | mb |
Short-period P-waves | M6.5 |
| Surface-wave magnitude | ms |
Long-period surface waves | M8.3 |
| Duration magnitude | md |
How long the shaking lasts on the record | Small events only |
| Moment magnitude | mww, mw |
The physical size of the rupture itself | Does not saturate |
Saturation happens because short-period waves stop getting stronger once the rupture is much larger than the wavelength being measured. Push a M8 and a M9 through an mb calculation and you get nearly the same answer.
The fix is moment magnitude, which is computed from the actual physics of the rupture rather than the height of a wiggle. But moment magnitude needs long-period waves that take longer to arrive and longer to process. For a significant earthquake, the authoritative mww solution typically lands 10 to 30 minutes after the event, well after the first headline number.
So the sequence you see on a tracker is often: a saturated fast estimate, then the real one.
The number in the first news alert is rarely the final number
For any earthquake above roughly M6.5, expect the initial figure to be provisional and quite possibly low. If a story quotes a magnitude within a few minutes of a major event, check the USGS event page again an hour later.
Depth Is the Least Reliable Number of All
Focal depth matters enormously, for reasons we unpack in why shallow earthquakes do more damage. A M6.0 at 8 km can devastate a town while a M6.0 at 200 km beneath the same town is barely felt. Yet depth is the parameter early solutions resolve worst.
The reason is geometric. Locating an earthquake horizontally is straightforward when stations surround it. Resolving how deep it was requires either a station almost directly above the source or specific wave phases that take time to identify. Without those, depth and origin time trade off against each other: a slightly deeper source with a slightly earlier origin time fits the same data almost as well.
When the data simply cannot constrain it, USGS fixes the depth at a default value, most often 10 km, and flags it. Other defaults appear too, including 5 km in mid-continental regions and on mid-ocean ridges. If you see exactly 10.0 km on a fresh event, that is often a placeholder rather than a measurement, and it frequently changes once better phases are analyzed. The value is not arbitrary, though: 10 km is roughly where well-constrained shallow earthquakes genuinely cluster, so plenty of events really do belong there.
Why USGS, EMSC, and JMA Report Different Numbers for the Same Quake
You will sometimes see the same earthquake listed as M7.1 on one site and M6.8 on another, at the same moment. Neither is wrong. Different agencies draw on different networks and publish independently:
- USGS (United States) runs the global National Earthquake Information Center feed.
- EMSC (Euro-Mediterranean) operates no stations of its own. It merges real-time solutions from around 65 European and Mediterranean networks, which is why it often carries regional events faster, and smaller ones, than the global USGS feed does.
- JMA (Japan) uses its own magnitude scale calibrated for Japanese seismicity.
- GFZ (Germany) runs the global GEOFON network and publishes its own rapid solutions, as do various national centers.
They see different subsets of stations, may prefer different magnitude scales, and review on different timelines. Agencies using the same scale usually settle to within about 0.2 to 0.3 of each other inside a day, which reflects the inherent uncertainty of the measurement rather than a disagreement waiting to be resolved. Where an agency uses a different scale entirely, as JMA does in Japan, the gap is permanent by design and never converges. WeatherAI lets you pick USGS or EMSC as your earthquake source, which is worth setting deliberately: if you are in or watching Europe, EMSC often reports smaller regional events that the global USGS feed does not carry at all.
Events Can Also Vanish
Revision runs in both directions, and occasionally an earthquake is deleted outright. Automatic detections get removed when review shows they were:
- Quarry or mine blasts, which are common and look superficially like small shallow earthquakes.
- False triggers from instrument glitches or a station reporting bad data.
- Duplicates of an event already in the catalog under a different ID.
- Phantom events created when waves from a large distant earthquake are misread as a local one, which is why a big quake is often followed by a brief flurry of spurious small ones worldwide.
If an event disappears from a live tracker between refreshes, this is usually why.
How Much Revision Is Normal?
A rough guide for what to expect after the initial post:
| Event Size | What Usually Happens |
|---|---|
| Under M4 | Small tweaks or deletion. Many are never human-reviewed at all. |
| M4 to M6 | Magnitude typically settles within about 0.2. Depth may move noticeably. |
| M6 to M7.5 | Expect a shift to a moment magnitude solution within roughly half an hour. Changes of 0.2 to 0.4 are common. |
| Above M7.5 | Expect the first figure to be low. Revisions of 0.5 or more, sometimes over several hours, are normal. |
What This Means When You Are Watching a Live Tracker
A few habits make live earthquake data much easier to read:
- Treat anything under an hour old as provisional, especially for large events.
- Check the magnitude type, not just the number. An
mwwlabel means a moment magnitude solution is in and the figure is fairly settled. Anmbormlon a large earthquake means you are looking at an early estimate. - Read depth alongside magnitude. A moderate shallow quake near people matters more than a larger deep one under open ocean.
- Watch felt reports. Public "Did You Feel It" submissions and ShakeMap intensity often describe real-world impact better than magnitude alone.
- Do not re-litigate the first number. The revision is the process working, not a mistake being fixed.
Our live earthquake tracker pulls the USGS feed directly and links every event to its official USGS page, which shows the reviewed solution as it evolves. For an event inside another network's territory, the local agency is the authority of record.
FAQ
Why did the earthquake magnitude go up?
Almost always because the first estimate was computed before the full rupture finished, or with a magnitude scale that saturates for large earthquakes. As long-period seismic waves reach more stations, a moment magnitude solution replaces the early estimate and is usually higher for major events.
Is the first reported magnitude wrong?
No. It is a preliminary estimate published deliberately fast, using the little data available in the first minute or two. Agencies accept that early figures will be refined because rapid notification is more valuable than waiting for a final answer.
How long until an earthquake magnitude is final?
A moment magnitude solution for a significant earthquake typically appears 10 to 30 minutes after the event, and a reviewed solution within hours to a day. Catalog values can still be revised months or years later during research reanalysis, though changes at that stage are small.
Why do USGS and EMSC show different magnitudes?
They are independent agencies with different seismic networks, different station coverage, and sometimes different preferred magnitude scales. Each publishes its own solution on its own review schedule, and the two usually converge within about a day.
Why is the depth always 10 km?
When the available data cannot resolve depth, USGS fixes it at a default of 10 km rather than publishing a meaningless number. An exact 10.0 km on a recent event is often a placeholder that will change once more wave phases are analyzed.
Why did an earthquake disappear from the list?
Automatic detections are deleted during review if they turn out to be quarry blasts, instrument glitches, duplicates, or phantom events created by misreading waves from a large distant earthquake.
Always Verify with Official Sources
Earthquake parameters are revised continuously and any live tracker is a mirror of an upstream feed at a moment in time. For authoritative event details, always use the USGS Earthquake Hazards Program, and for protective action follow your local emergency management agency and official tsunami warning centers.
Follow Earthquakes with WeatherAI
WeatherAI keeps seismic activity alongside your weather, on iPhone, iPad, Mac, and Vision Pro:
- Nearby and global earthquake notifications with magnitude thresholds you set, so small events stay quiet
- Choose your data source, USGS or EMSC, depending on where you are watching
- Magnitude and distance filtering on maps and lists, plus full event detail
- Tsunami bulletins surfaced alongside earthquake information
- Earthquake widgets for your Home Screen and Lock Screen with a minimum magnitude you control
WeatherAI reports earthquakes that have already been detected. It is monitoring, not earthquake early warning, and it does not replace official emergency alerts.
Download WeatherAI for iPhone →
Keep reading: Richter vs moment magnitude explains why the scale is logarithmic, earthquake depth explained covers why shallow quakes do more damage, and foreshocks, mainshocks and aftershocks covers what happens next. You can also browse the live earthquake tracker and our earthquake glossary terms.
Keep Reading
These articles cover adjacent weather questions, practical explainers, and planning workflows related to this topic.
Weather Education
Richter vs Moment Magnitude: Why M7 Is Not Twice M3.5
The Richter scale was retired for large earthquakes decades ago. Here is how moment magnitude works, why each whole number means about 32 times more energy, and what magnitude does and does not tell you about damage.
Weather Education
Earthquake Depth Explained: Why Shallow Quakes Do More Damage
A magnitude 6.7 killed 57 people. A magnitude 6.8 three times deeper killed no one directly. Here is how earthquake depth controls shaking, why deep quakes are felt farther but weaker, and when depth stops mattering.
Weather Safety
Foreshocks, Mainshocks, and Aftershocks: What to Expect After a Large Earthquake
Aftershocks get rarer over time, not smaller. Here is what the labels actually mean, why a foreshock can only be identified after the fact, how long sequences last, and what to do while one is going on.
Useful Next Stops
Jump from the explainer into tools, feature pages, and seasonal hubs that deepen the topic.
Guide
Earthquake Depth Explained
Why a shallow quake shakes far harder than a deeper one of the same magnitude, and when depth stops mattering.
Guide
Richter vs Moment Magnitude
How moment magnitude replaced the Richter scale, and why each whole number means about 32 times more energy.
Guide
Foreshocks, Mainshocks, and Aftershocks
What the labels really mean, how long sequences last, and why aftershocks get rarer over time but not smaller.
Glossary
Aftershock
A smaller earthquake following a mainshock in the same area, caused by the crust adjusting to redistributed stress.