Basin-crossing tsunamis are modelled with the shallow-water equations —
valid whenever a wave's length is far greater than the water depth beneath it, which is
exactly the case for a tsunami crossing an ocean basin. This scene solves a linearised
form of those equations on a finite-difference grid: a leapfrog update advances the sea
surface elevation η using a Laplacian weighted by local water depth
H(x,y), the same core idea used in production codes like GeoClaw or MOST,
just at a scale a browser can animate in real time.
c = √(g·H), so speed falls as depth falls. Offshore the wave crosses at hundreds of km/h; on the shelf it slows sharply.Real-time tsunami forecasting systems such as NOAA's MOST/ComMIT and GeoClaw run thousands of scenario simulations offline, then interpolate a forecast within minutes of an earthquake — GPU acceleration and adaptive mesh refinement are what make it possible to resolve metre-scale coastal detail without simulating the whole Pacific at that resolution.
A live shallow-water finite-difference solver: trigger a seafloor displacement and watch the resulting wave cross an open basin, shoal as the seabed shallows, and run up a coastline whose slope you control.
Wave speed c = √(g·H) falls as depth falls, forcing the wave to slow and stack up (shoal) near shore — the same mechanism that turns a barely-visible open-ocean tsunami into a destructive coastal wave.
Set the seafloor displacement and ocean depth, choose a steep or gradual coastal slope, and trigger the tsunami. Toggle the adaptive-mesh overlay to see why coastal forecasting codes refine resolution only where the seabed changes fastest.
Operational systems like NOAA's MOST/ComMIT pre-compute thousands of scenarios offline so that, after a real earthquake, a forecast can be assembled by interpolation within minutes rather than by running a fresh simulation from scratch.