Two earthquakes of ML = 5.4 occurred within half an hour of each other within the Hilea area of Southern Hawaii on 21 January 1982. The aftershock distribution suggests that together they ruptured an area of approximately 5-km radius, and their joint equivalent magnitude was 5.6. The first motions indicate faulting on a near horizontal plane at 10 km depth, with the crust slipping to the southeast.
The seismicity rate in the source area was studied using the earthquake catalog of the Hawaiian Volcano Observatory. This catalog contains some reporting rate changes that affect the count of smaller earthquakes strongly, with a period of low reporting during 1974 to 1977. Although we have used 2.5 as the minimum magnitude of homogeneous reporting after 1971, some of the artificial rate changes are still present in the data. The catalog was declustered using Reasenberg's algorithm, and a magnitude correction of −0.1 was applied to the data between 1974 and 1984.
The seismicity rate for the period of November 1971 through 1985 was examined in four adjacent regions; one of these contained the aftershocks. The aftershock volume and the 3-km annulus around it showed a period of 46 weeks of low seismicity rate immediately before January 1982, during which the rate was decreased by 87 per cent. The seismicity rate in the other three volumes was normal during this time. We conclude that this low reporting rate was not likely due to artificial changes in the catalog. The fact that the quiescence anomaly coincided in space with the 1982 aftershock volume and that its termination coincided with the 1982 main shocks suggests that the quiescence was a precursor.
Comparing the seismicity rate within all possible 46-week windows to the background rate, we found that the precursory rate decrease was more significant than any other rate decrease in all volumes studied except artificial low rate periods. Thus this quiescence precursor could be recognized without false alarms. However, the statistical significance estimated by the z- and β-tests was low, ≦75 per cent and ≦49 per cent, respectively. More case histories are needed to determine empirically the thresholds of these tests for accepting precursory anomalies without too many false alarms.
The M = 6.6 Kaoiki earthquake that was located about 25-km north of the 1982 source area was preceded by 125 weeks of quiescence, while the M = 7.2 Kalapana earthquake quiescence precursor lasted about 200 weeks. These observations suggest that in Hawaii quiescence precursor times may be a function of magnitude.