Man-made changes of seismicity rates

Abstract

Seismicity catalogs contain important information about processes which occur in seismically active regions of the earth. Many authors have examined these catalogs for patterns and variations in patterns which might reflect changes in these processes. We have found that these catalogs include a complex mixture of real and man-made changes. One must identify and account for the man-made changes before the real ones can be identified and understood. Many man-made changes in seismicity catalogs are manifested as changes in seismicity rates and can, therefore, be identified through careful examination of these rates. Obvious effects include increases or decreases in the detection and reporting of smaller events which accompany the installation or closure of seismic stations. These types of changes can be recognized by examining the distribution of seismicity rate changes in the magnitude domain. They are characterized by increases or decreases in the number of small events in the catalog at times when the number of larger events remains constant. Systematic changes in the magnitudes assigned to events can also be identified by examining seismicity rates because they cause apparent changes in rates of data sets with magnitude cutoffs. The sign of the apparent rate change depends on the sign of the magnitude change and the type of cutoff used. The effects of detection changes can be easily remedied by using a magnitude cutoff, which eliminates the smaller events from consideration. The techniques we have developed allow one to determine the optimum cutoff. Magnitude corrections are necessary for remedying the effects of systematic changes in magnitude estimates. These corrections can be determined through modeling of observed rate changes caused by these shifts. Man-made changes are present in all seismicity catalogs whether local, regional, or teleseismic. They must be accounted for if these catalogs are to provide meaningful information on real process changes in the earth.