AfterQuake estimates further shaking using public USGS earthquake data. It cannot inspect a building or give an all-clear. Early data can change. Check the selected date when exploring past earthquakes.
How the forecast at your spot works
1. How many. First, we fit a Reasenberg–Jones aftershock model to events at or before your selected time. This prototype uses one prior everywhere, while USGS operational forecasts use regional parameters and more detailed handling of missing events.
2. Where. We then use past epicentres as proxies for future source locations, with a 6 km distance smoothing term and a fixed 10 km depth. Together, these form a simplified spatial model.
3. How hard. To estimate shaking at your spot, we use the Allen–Wald–Worden (2012) active-crust intensity equation, including its scatter, and integrate over a normalised M5–8.5 magnitude distribution. This leaves out smaller aftershocks and sources beyond 400 km.
4. Your ground. We also account for local ground conditions with a soil term fitted to ShakeMaps: ΔMMI = × log10(Vs30/760).
5. Your building. We choose a damage threshold from the Modified Mercalli scale or, for "typical local building", from the country's USGS PAGER casualty model. To represent possible weakening, the threshold drops one level if mainshock shaking here was near that level, or if the ground-failure model gives at least 10% liquefaction.
6. Time period. For your chosen stay, we integrate the aftershock rate over the whole 5-minute to 24-hour interval. We then project when the chance during that interval falls below your limit, assuming the model continues unchanged.
7. Live earthquakes. For quakes opened from the globe, we use the same model with the USGS catalogue, ShakeMap and PAGER report. All three are checked for revisions every 2 minutes while the quake is open. Ground-failure maps are not loaded, and because early ShakeMaps are often automatic, later revisions can change the estimate. This is not an official USGS forecast.
Replay limits (case studies). Although the timeline filters earthquake origin times, the catalogue contains later revisions, the mainshock ShakeMap may have been processed years later, and the soil coefficient was fitted using these sequences. The replay therefore illustrates past sequences; it cannot reconstruct a real-time forecast or independently validate local shaking.
Using the result. These are prototype estimates, with local detail limited by the input grids. Whether a building is safe to enter still needs a structural inspection and local authority guidance.