The Parkfield earthquake prediction is generally stated as a 95% probability that the next moderate earthquake there should occur before January 1993. That time limit is based on a two-sided 95% confidence interval. Because at the time of the prediction (1985) it was already clear that the earthquake had not occurred prior to 1985, a one-sided 95% confidence interval would have been more appropriate. That confidence interval ended in October 1991. The Parkfield prediction was based on an extrapolation of five of the six events in the 1857 to 1966 earthquake sequence; the 1934 event was omitted because it did not fit the regularity exhibited by the other data. The fallacy in the prediction is that it did not take account of other less-contrived explanations of the Parkfield seismicity (e.g., not excluding the 1934 event). Even if the Parkfield earthquake should occur in the near future, it would be better explained by less-contrived hypotheses.
[1]
Thomas V. McEvilly,et al.
Recurrence models and Parkfield, California, earthquakes
,
1984
.
[2]
M. Wyss,et al.
Decrease in deformation rate as a possible precursor to the next Parkfield earthquake
,
1990,
Nature.
[3]
R. E. Habermann,et al.
Seismic quiescence at Parkfield: an independent indication of an imminent earthquake
,
1990,
Nature.
[4]
W. H. Bakun,et al.
The Parkfield, California, Earthquake Prediction Experiment
,
1985,
Science.
[5]
R. Plackett,et al.
Introduction to Statistical Analysis.
,
1952
.
[6]
J. C. Savage.
Criticism of some forecasts of the national earthquake prediction evaluation council
,
1991,
Bulletin of the Seismological Society of America.
[7]
H. Jeffreys,et al.
The Theory of Probability
,
1896
.