Aftershock Forecasts: Probabilities After a Large Earthquake
Aftershock forecasts use statistical patterns from prior sequences and observations from the current sequence to estimate expected counts and the chance of at least one earthquake above selected magnitude thresholds during stated time windows.
Timeline
- 2026-10-13: Current USGS operational aftershock-forecast documentation, earthquake-sequence guidance and FEMA safety material was reviewed for this guide.
- 2026-10-13: The current authoritative sources cited in this guide were reviewed.
An aftershock is an earthquake in the same general area after the largest event so far in a mainshock-aftershock sequence. The mainshock changes stress on and around the ruptured fault, and nearby faults or unruptured patches adjust through additional earthquakes. “Mainshock,” “aftershock” and “foreshock” are relative labels, not different physical species. If a later earthquake is larger, it becomes the mainshock and the earlier event is relabeled a foreshock. That possibility is one reason an official forecast includes the chance of an earthquake larger than the first event. [1][2][3]
Aftershock sequences show repeatable group patterns even though an individual earthquake’s exact time, place and magnitude cannot be predicted. Larger mainshocks generally produce more aftershocks over a larger region and longer period. Smaller aftershocks greatly outnumber larger ones. Most importantly, the rate usually falls with elapsed time, roughly inversely with time in a typical sequence: USGS illustrates that the first day may have about ten times as many aftershocks as the tenth. The declining rate reduces near-term likelihood, but it does not make the possible magnitudes shrink with time. [1][3][4]
An operational forecast turns those population patterns into probabilities and expected counts. USGS reports the likely number of earthquakes above selected smaller magnitude thresholds and the probability of at least one event above larger thresholds that could damage. Results are presented for windows such as the next day, week, month and year. A probability always belongs to a particular magnitude threshold, geographic region and time interval. “Twenty percent” without those qualifiers is incomplete, and it does not mean shaking will occupy twenty percent of the time or area. [1]
The first forecast can be produced soon after the mainshock using its magnitude and generic parameters learned from past sequences in comparable tectonic settings. USGS waits for the preferred magnitude to stabilize before issuing it because even a modest magnitude revision changes expected productivity. As instruments record the actual sequence, the model can blend the generic expectation with sequence-specific behavior. Forecasts are updated because rates evolve and a substantial aftershock can temporarily raise subsequent activity. The issue time and planned next update therefore matter as much as the percentage displayed. [1][5]
A falling daily rate and a growing cumulative probability are compatible. Tomorrow may be less active than today, yet a one-month window includes many more opportunities than a one-day window. For the same magnitude threshold and starting time, a longer interval will generally have at least as much probability because it contains the shorter interval plus additional days. Similarly, the chance of at least one M4-or-larger event is greater than the chance of at least one M6-or-larger event because the lower threshold includes more possible earthquakes. Compare like thresholds and windows before describing a trend. [1][4]
The forecast area is broad situational guidance, not a map of the next epicenter. Aftershocks cluster near the mainshock rupture and where activity has already occurred, commonly within a few fault lengths, but there is no perfectly sharp physical boundary. USGS operational pages currently show the region to which the probabilities apply and the earthquakes recorded so far; they do not calculate a point-by-point spatial probability map within that region. Dense seismic networks also detect smaller and more distant events that a sparse network might miss, affecting the catalog used to interpret the sequence. [1][3][6]
Forecasting is not earthquake prediction or early warning. A prediction would specify the date and time, location and magnitude of a future earthquake, which science cannot do reliably. An aftershock forecast gives conditional probabilities across windows based on statistical behavior. Earthquake early warning starts only after rupture begins and may provide seconds of notice before strong shaking reaches another location. A sequence product describes earthquakes that have happened and how they cluster; an aftershock forecast estimates what is likely next. Keeping those products separate prevents a probability update from being mistaken for an alarm. [5][7]
Use the forecast to maintain readiness, not to decide that a damaged structure is safe. Aftershocks can produce additional damage, landslides or falling debris even when smaller than the mainshock. Follow local officials and qualified building inspectors, stay out of visibly damaged buildings, expect utilities and roads to be affected and keep alerts active. If shaking begins, Drop, Cover and Hold On; protect the head and neck and do not run through falling hazards. In a tsunami-prone coastal area, follow evacuation guidance after strong or long shaking without waiting for an aftershock percentage. [1][8]
Sources
- U.S. Geological Survey — Aftershock Forecast Overview
- U.S. Geological Survey — Foreshocks, Aftershocks: What’s the Difference?
- U.S. Geological Survey — Earthquake Sequence Product
- U.S. Geological Survey — Aftershocks, Swarms and What They Mean
- U.S. Geological Survey — Early Warning, Forecasts, Probabilities and Prediction
- U.S. Geological Survey — Some Facts About Aftershocks to Large Earthquakes in California
- U.S. Geological Survey — What Is the Probability an Earthquake Is a Foreshock?
- FEMA Ready.gov — Hazard Information: Earthquakes