Why are so many earthquake reports exactly 10.0 km deep? A learner scrolls through an earthquake catalogue and notices several events with a depth of 10 km. The number looks wonderfully precise. “Ten kilometres exactly,” she says. “The instruments must have measured the focus to the nearest hundred metres.”
Sometimes that conclusion is wrong. The U.S. Geological Survey explains that when available seismic data cannot determine an earthquake’s depth reliably, a catalogue system may assign a fixed depth. In many areas, 10 km is used as a reasonable default because many crustal earthquakes occur near that depth. A displayed 10 km can therefore be a solution constraint rather than an exact direct measurement.
Quick answer
No. An earthquake report that says “depth = 10 km” does not automatically mean scientists measured the earthquake at exactly 10.0 km. Some 10 km depths are well constrained by seismic observations; others are fixed or assigned because the available data cannot solve depth reliably. The learner job is to inspect the metadata, uncertainty and solution method before deciding what the number means.
The useful habit is broader than earthquakes: a precise-looking number can be a model input, constraint, default or reporting convention rather than a precise observation.
The owned learner job — and what belongs elsewhere
This Reality Lab owns one evidence-transfer problem: how to read a scientific earthquake catalogue when a depth value looks measured but may be fixed by the location procedure. It does not replace earthquake physics, seismic-wave mechanisms, triangulation, coordinate systems, measurement uncertainty or model fitting as standalone science owners.
For the distinction between observation and inference, route to How to Tell Observation, Inference, Prediction and Explanation Apart in PSLE Science. For the separate issue of treating an epicentre dot as an exact location, see PSLE Science Reality Lab Vol No.167 | “The Epicentre Dot Is Right Here” — Is the Earthquake Location Exact to the Dot?.
Original report card: one number, two possible meanings
Consider this invented earthquake catalogue entry:
Event EK-478
Origin time: 14:21:08 UTC
Latitude: 35.20°
Longitude: 117.80°
Depth: 10.0 km
Magnitude: 3.4
Status: preliminary
Depth note: fixed during automatic solution
The depth field contains a number, but the note changes how we must read it. “10.0 km” is part of the current earthquake-location solution. It is not evidence that a drill, camera or ruler reached the earthquake and measured exactly ten kilometres.
Read the metadata before trusting the decimal places
A scientific database often separates the value from information about how the value was produced. That surrounding information is metadata. For an earthquake depth, useful metadata can include uncertainty, location method, number and geometry of seismic observations, whether the solution was automatic or reviewed, and whether depth was allowed to vary.
If you copy only “10.0 km” and remove the metadata, you may make the result sound more certain than the catalogue intended.
Observed, calculated, constrained and reported
| Layer | Example | Possible mistake |
|---|---|---|
| Observed | Seismometers record wave arrivals and ground motion at stations. | Saying a station directly measured the underground depth. |
| Calculated | A location algorithm estimates origin time and position from seismic observations and a model. | Treating every solved parameter as equally well constrained. |
| Constrained | Depth may be fixed to a chosen value when data do not constrain it reliably. | Reading the fixed value as a direct measurement. |
| Reported | The catalogue displays latitude, longitude, depth and other fields. | Ignoring uncertainty and method because the table looks exact. |
The earthquake is real. The seismic recordings are real. The catalogue entry is real. The scientific question is what kind of evidence each field represents.
Why depth can be harder to constrain than horizontal location
USGS documentation describes depth as commonly the least well-constrained part of an earthquake location. One reason is geometry. Seismic stations sit mainly at Earth’s surface, so their observations may constrain horizontal position better than vertical position, especially when the nearest stations are far from the earthquake.
Different combinations of depth, origin time and model assumptions can sometimes fit the available wave-arrival observations similarly. If the data do not strongly distinguish among those possibilities, allowing depth to wander freely can create an unstable or unrealistic solution.
What a fixed depth does
A fixed depth is a constraint placed on the location solution. Instead of asking the algorithm to solve latitude, longitude, origin time and depth all at once, the system holds depth at a chosen value while solving the other parameters.
This can make the solution more stable when depth is poorly constrained. But stability is not the same as proof that the fixed depth equals the earthquake’s exact physical depth.
Why 10 km appears so often
USGS explicitly notes that many earthquakes are shown at 10 km because depth was fixed when available data were too poor to compute a reliable depth. Ten kilometres is a reasonable starting or fixed value for many shallow crustal settings. Other networks or tectonic settings may use other fixed depths.
Therefore, the repeated number itself is a clue to investigate the method. Repetition does not prove that every event actually occurred on one perfect underground sheet ten kilometres below the surface.
A number can be useful without being an exact measurement
Suppose a preliminary catalogue needs a stable horizontal location quickly. A fixed 10 km depth can allow a useful preliminary solution even if later analysis finds that the earthquake was 7 km, 14 km or another depth.
Calling the initial value “fixed” does not make the catalogue dishonest. It makes the modelling choice visible. Scientific integrity improves when assumptions and constraints are documented.
Worked case 1: 10.0 km becomes 7.8 km
An automatic solution for an invented earthquake initially reports depth 10.0 km with the depth fixed. Several hours later, more seismic arrivals are reviewed and the catalogue reports 7.8 km with a depth uncertainty estimate.
Did the earthquake rise 2.2 km after it happened? No. The physical event is in the past. The estimate changed because the evidence and solution improved. This is the same broad habit used elsewhere in Reality Lab: a changing scientific record can reflect changing knowledge rather than a changing past event.
Worked case 2: two earthquakes both listed at 10 km
Earthquake A and Earthquake B are hundreds of kilometres apart. Both catalogue entries say 10.0 km. A student says, “They occurred at exactly the same depth.”
Before accepting that claim, check whether either depth was fixed. If both were assigned the same default, the equality may come from the processing rule rather than from identical physical depths.
Worked case 3: the very precise display
A website displays “Depth: 10.00 km”. Two decimal places make the number look precise to ten metres. But formatting precision is not measurement certainty. The database may simply be printing every depth field with the same number of decimal places.
Never infer uncertainty from the number of digits shown unless the reporting system explicitly says those digits represent measurement precision.
Worked case 4: depth error is larger than the difference
One event is reported at 8 km depth and another at 11 km. A headline says, “The second earthquake was exactly 3 km deeper.” If both depth estimates have large uncertainties, the exact 3 km comparison is too strong.
USGS ComCat provides a depthError field for uncertainty in the reported depth. A difference between central values should be interpreted alongside that uncertainty, not treated as infinitely exact.
Worked case 5: automatic does not mean wrong
A social post says, “The depth was automatic, so the whole earthquake report is fake.” That is another overreach. Automatic processing can rapidly turn real instrument observations into preliminary scientific products. The correct response is to recognise the preliminary status and check later review, not to reject all automatic results.
Scientific scepticism is disciplined calibration of confidence, not automatic disbelief.
Worked case 6: reviewed does not mean infinitely exact
A reviewed event later receives a revised depth. A student is surprised: “But it was already reviewed.” Review can improve a result without making uncertainty vanish. New stations, waveform analysis or improved models can support later revisions.
The status of a scientific product tells you something about its processing history; it is not a magic label that removes all uncertainty.
Worked case 7: the map cross-section trap
A cross-section plots many earthquakes. A horizontal row appears at exactly 10 km depth. The viewer concludes that a geological layer at 10 km is causing all those earthquakes.
That pattern could be real, but first check whether many depths were fixed to 10 km. A processing convention can create a visual horizontal band. Representation patterns need provenance before they become geological explanations.
Worked case 8: one number in a news headline
A news-style headline says, “Earthquake struck exactly 10 km below the town.” The catalogue actually lists a fixed preliminary depth of 10 km and an epicentre several kilometres from the town centre.
Two transformations happened: an assigned depth became “exactly”, and an area reference became “below the town”. A strong learner checks both horizontal and vertical evidence before accepting the sentence.
The precision ladder
- Displayed number: what appears in the table.
- Method status: fixed, freely solved, automatic, reviewed or otherwise documented.
- Uncertainty: how tightly the evidence constrains the value.
- Independent support: whether additional observations agree.
- Claim scope: how precisely the result can be communicated.
Do not jump from Step 1 directly to Step 5.
Representation check: catalogues look more certain than they feel
Tables need one value in each cell. A catalogue cannot print a long essay inside the depth column. So uncertainty, method and status appear in neighbouring fields, metadata pages or documentation.
The compact table is not misleading by itself. The mistake is reading one cell as the entire evidence story.
Comparison check: before ranking earthquakes by depth
- Were both depths solved using comparable methods?
- Was either depth fixed?
- What are the depth uncertainties?
- Were the events reviewed to similar levels?
- Are the station networks similarly dense?
- Are the earthquakes in comparable tectonic settings?
- Are the catalogue versions from the same time?
A clean numerical ranking can be scientifically weak if the underlying evidence quality differs.
Method check: what would help constrain depth better?
- Nearby seismic stations.
- Good station geometry around the event.
- Clear seismic-wave arrival times.
- Appropriate Earth-velocity models.
- Additional waveform information.
- Human review of questionable arrivals.
- Independent networks or later analyses where available.
These are examples of evidence that can improve a location solution. They are not a PSLE requirement to calculate earthquake depth; the Reality Lab job is simply to recognise why the reported number may have limits.
Alternative explanations for a row of 10 km earthquakes
- Many earthquakes really occurred near 10 km.
- Some depths were fixed at 10 km.
- The catalogue rounded nearby depths.
- A subset came from automatic preliminary solutions.
- A plotting or filtering choice emphasised one value.
The visible pattern does not decide among these explanations on its own. Check the metadata.
What strengthens a claim that an earthquake truly occurred near the reported depth?
- The depth was freely solved rather than fixed.
- Depth uncertainty is small relative to the claim.
- Nearby stations constrain the solution well.
- Reviewed solutions remain similar.
- Independent analyses support a comparable depth.
- The catalogue documents a suitable location method.
What weakens an exact-depth claim?
- The depth is explicitly fixed.
- The event is preliminary and automatic.
- Depth uncertainty is large or absent.
- Few nearby stations are available.
- Later catalogue versions revise the depth substantially.
- The claim relies only on decimal places shown in a table.
How far can the conclusion travel?
If the catalogue says the depth was fixed at 10 km, a careful sentence is: “The preliminary location solution used a fixed depth of 10 km.” That preserves the scientific status of the number.
If a reviewed solution later reports 8.2 km with a justified uncertainty, a stronger depth claim may be possible. Even then, communicate the value as an estimate with uncertainty rather than an infinitely exact underground coordinate.
Tempting reasoning that fails
- “10.0 km means measured to 0.1 km.” Display precision is not uncertainty.
- “Every 10 km event happened at exactly the same depth.” Some may share a fixed default.
- “Fixed means fake.” A documented constraint can be a legitimate scientific solution choice.
- “Reviewed means exact forever.” Review improves evidence but does not abolish uncertainty.
- “A later revision means the first event changed.” The estimate changed, not the past earthquake.
Original PSLE-style transfer case: three catalogue entries
This is original practice, not a copyrighted examination question.
| Event | Depth | Metadata |
|---|---|---|
| A | 10.0 km | Automatic; depth fixed |
| B | 9.6 km | Reviewed; depth error 1.2 km |
| C | 10.0 km | Reviewed; depth solved; depth error 0.8 km |
Question 1: Which 10 km value should not be interpreted as a directly solved depth?
Explained answer: Event A, because its metadata state that depth was fixed.
Question 2: Do Events A and C prove two earthquakes occurred at exactly the same physical depth?
Explained answer: No. Event A used a fixed constraint, while Event C has a solved estimate with uncertainty.
Question 3: Why should Event B not be described as “exactly 9.6 km deep”?
Explained answer: The catalogue provides depth uncertainty, so 9.6 km is an estimate rather than an infinitely exact value.
Question 4: What is the strongest safe sentence about Event A?
Explained answer: “The automatic location solution used a fixed depth of 10 km.”
Delayed independent return
- Why can a catalogue show 10.0 km even when depth is poorly constrained?
- What is the difference between a fixed depth and a freely solved depth?
- Why do decimal places not reveal uncertainty automatically?
- What does a depth-error field add?
- Why might a row of earthquakes at 10 km be partly a processing pattern?
Self-check: A system can fix depth to stabilise a solution; fixed means constrained rather than independently solved; formatting and uncertainty are different; depth error communicates how tightly the value is constrained; repeated defaults can create artificial visual bands.
Explained practice: value or meaning?
- “Depth: 10 km; fixed.” Treat 10 km as a location constraint in that solution.
- “Depth: 6.4 km; uncertainty ±0.7 km.” Treat 6.4 km as an estimate with stated uncertainty.
- “Depth: 10.00 km” with no metadata visible. Do not infer hundredths-of-a-kilometre certainty; seek the metadata.
- “Depth revised from 10 km to 14 km.” The scientific estimate changed; the historical earthquake did not move.
Parent and tutor teaching guide: the fixed-number experiment
Hide a small object under one of several stacked books. Give the learner two clues that locate it horizontally but do not reveal which book layer contains it. Ask them to produce a stable diagram by temporarily fixing the layer at “Book 3”. Label the diagram clearly: layer fixed.
Then provide a new clue that indicates Book 2. The learner should update the model without saying the object physically moved. This makes the distinction between changing object and changing estimate tangible.
For three students, assign roles: instrument reader, who sees the observations; solver, who must sometimes fix one parameter; and claim checker, who decides how the result can be described. Rotate roles and change which parameter is fixed.
Authoritative sources and curriculum frame
- SEAB — 2026 PSLE Science syllabus: current assessment includes interpreting and analysing information, evaluating observations, information and methods, and communicating explanations and reasoning.
- MOE — 2023 Primary Science Teaching and Learning Syllabus: supports healthy scepticism, evidence evaluation, assumptions and uncertainty.
- USGS FAQ — Why do so many earthquakes occur at a depth of 10 km?: explains that 10 km may be a fixed depth assigned when available data cannot determine depth reliably.
- USGS — Search Earthquake Catalog / ComCat documentation: documents earthquake catalogue fields, including depth uncertainty and event-solution metadata.
The quiet habit to keep
When a scientific table gives you a neat number, ask one question before doing arithmetic with it: What kind of number is this? Measured, calculated, modelled, fixed, rounded and estimated numbers can all be useful — but only when their origin travels with their meaning.
