Passive data

To understand the sources captured in a passive seismic survey, I hoped planewave decomposition/inversion could help analyze the data. However, unless identifiable events are present, analysis of passive data before cross-correlation does not produce interpretable results. After correlation, the unique character of the individual sources is lost. In effect, transforms applied to the data before correlation simply reshuffle the randomness apparent in the raw traces. The passively collected solar data Rickett and Claerbout (1999) was analyzed to prove this failure. Figure  shows the raw solar data and its autocorrelation. Clearly, there are events to be found within the raw data that are masked before correlation. Figure  shows the linear Radon domain inversions for data defined by the panels of Figure . Because the correlated data is radially symmetric from the center and only has events in the upper third of the time axis, its model space is much smaller, though sampling between the two is the same.

 

sun.dat
Figure 5 Passive seismic data from the sun and its autocorrelation.

 

sun.mod
Figure 6 Model space produced with 20 iterations of planewave decomposition inversion from the data shown in Figure .