Seismic simulations are used for example for geothermal and mining exploration, and for studying earthquakes. Current seismic simulations usually manually define a refinement area around the source and the regions of interest. Recently an adaptive mesh-refinement technique is presented for tsunami simulations. It uses the adjoint equation to run the simulation in reverse from a point of interest. The result is combined with a forward simulation to construct the refinement. We adapt this method in this thesis for elastic wave equations. We primarily use statically adaptive mesh refinement. The generated refinements follow the paths which the waves take to the receivers. This results in lower error than manually created refinements for a test scenario which produces only P-Waves and a scenario which models the geology under Helsinki as a 1d-velocity model. However, the manual refinement is better in a simpler system which only consists of two different materials. A prototype for creating dynamically adaptive grid with the same method was also implemented and tested with promising results.     «

Seismic simulations are used for example for geothermal and mining exploration, and for studying earthquakes. Current seismic simulations usually manually define a refinement area around the source and the regions of interest. Recently an adaptive mesh-refinement technique is presented for tsunami simulations. It uses the adjoint equation to run the simulation in reverse from a point of interest. The result is combined with a forward simulation to construct the refinement. We adapt this met...     »