Changes in the rate of occurrence of smaller events have been recognized in the rupture zones of upcoming large earthquakes in several postearthquake and one preearthquake study. A data set in which a constant portion of the events in any magnitude band are consistently reported through time is crucial for the recognition of seismicity rate changes which are real (related to some process change in the earth). Such a data set is termed a homogeneous data set.
The consistency of reporting of earthquakes in the NOAA Hypocenter Data File (HDF) since 1963 is evaluated by examining the cumulative number of events reported as a function of time for the entire world in eight magnitude bands. It is assumed that the rate of occurrence of events in the entire world is roughly constant on the time scale examined here because of the great size of the worldwide earthquake production system.
The rate of reporting of events with magnitudes above mb = 4.5 has been constant or increasing since 1963. Significant decreases in the number of events reported per month in the magnitude bands below mb = 4.4 occurred during 1968 and 1976. These decreases are interpreted as indications of decreases in detection of events for two reasons. First, they occur at times of constant rates of occurrence and reporting of larger events. Second, the decrease during the late 1960's has also been recognized in the teleseismic data reported by the International Seismological Centre (ISC). This suggests that the decrease in the number of small events reported was related to facets of the earthquake reporting system which the ISC and NOAA share. The most obvious candidate is the detection system.
During 1968, detection decreased in the United States, Central and South America, and portions of the South Pacific. This decrease is probably due to the closure of the VELA arrays, BMO, TFO, CPO, UBO, and WMO. During 1976, detection decreased in most of the seismically active regions of the western hemisphere, as well as in the region between Kamchatka and Guam. The cause of this detection decrease is unclear.
These detection decreases seriously affect the amount of homogeneous background period available for the study of teleseismic seismicity rate changes. If events below the minimum magnitude of homogeneity are eliminated from the teleseismic data sets the resulting small numbers of events render many regions unsuitable for study. Many authors have reported seismicity rate decreases as possible precursors to great earthquakes. Few of these authors have considered detection decreases as possible explanations for their results. This analysis indicates that such considerations cannot be avoided in studies of teleseismic data.