Not only historically, but even today, one of the most important problems in seismology is the location of earthquake sources: only a minority of all quakes is observed by eye-witnesses. On earth, thousands of seismometers have been installed since the dawn of instrumental seismology in the late 19th century. These allow both locating the earthquake focus and the determination of the uncertainty of this location.
The situation is more difficult for seismic experiments on other planets: the determination of the four source coordinates (three spatial coordinates plus time) requires – at least – a matching number of independent observations. If only few seismometers are available, classical location methods tend to under-estimate the uncertainty of the results. At the same time, arrival time determination on seismograms from the Moon, for example, suffers from the small moonquake magnitudes, hence the input for the location procedure already carries a significant uncertainty.
The LOCSMITH software (Knapmeyer, 2008), written in the MatLab language, implements a novel criterion for the optimization of solutions. It does not focus on searching the best fitting solution – which, when using insufficient data, is determined by input uncertainty rather than by the true position – but attempts to identify all locations that are compatible with the observations. The result is not a best fit solution which may be quite wrong, but a set of candidates that certainly contains the true solution.
Contact: Dr. Martin Knapmeyer