B L N Kennett
Articles written in Journal of Earth System Science
Volume 104 Issue 4 December 1995 pp 539-553
In the endeavour to achieve the goal of understanding the structure of the Earth in three dimensions, the limitations of the seismological probe for Earth structure have to be recognised. In geo-seismology there are high quality observations of a wide range of information on the seismic wavefield but only at a limited number of observation points around the globe. In contrast for the Sun, the other body for which seismological techniques have been routinely applied, there is the equivalent of a very high concentration of seismometers on one-hemisphere but very limited data quality. The restriction imposed by data quality is easily recognised but the restrictions imposed by the limited sampling of the globe are more subtle and have to be borne in mind in the interpretation of any images of seismic structure.
A wide range of studies has demonstrated the presence of heterogeneity in Earth-structure on a wide range of scales from planet-wide variations in the degree 2 spherical harmonics to the intensity of variations on the micro-scale in intensely deformed rock belts. Whatever scale of structure is being examined, the potential influence of other classes of heterogeneity has to be recognised and so an effort made to minimise the transfer of information from different scales which may contaminate the picture.
An important class of questions which must be faced in the development of threedimensional models is the nature of the reference model (or group of models) from which the model is to be derived, and also whether it is likely that limitations in the starting state will map into final images of structure. Further, are the various classes of data being used, compatible within the class of models which is being imposed?