Post-Tohoku Earthquake Probabilities for Tokyo

06-Feb-12

EQECAT is aware of recent press citing a 70% chance that Tokyo will experience a M7 earthquake in the next four years, attributed to Japanese researcher Professor Naoshi Hirata of the University of Tokyo's Earthquake Research Institute (ERI). There are significant uncertainties associated with this probability.

In a statement updated as of January 31, 2012 ERI clarifies that the increased earthquake probability in the vicinity of Tokyo is based on the observation of a 5-fold increase in the frequency of M3 - M6 earthquakes, which are largely assumed to be aftershocks, since the 2011 Tohoku earthquake. This increased frequency was extrapolated to M7 earthquakes using a Gutenberg-Richter relationship, which assumes unchanging earthquake frequencies over time.

On the contrary, however, the rate of aftershock activity is known to decrease rapidly with time, according to Omori's law. Thus, Gutenberg-Richter calculations of aftershock probabilities are relevant only over a short window of time. Given that this estimate was first made public in September, probability calculations based on current rates of aftershocks would already be considerably lower, as confirmed in the updated statement by ERI. In addition, it is difficult to differentiate triggered aftershock events from ongoing, natural background seismicity. ERI acknowledges in their statement that no attempt was made to separate these types of earthquakes.

A probability calculation based on aftershock occurrence is not indicative of medium- and longer-term earthquake risk, nor does it represent the standard of practice for earthquake risk modeling. In addition, the numerous uncertainties associated with a probability estimate based on aftershocks are very large, as acknowledged by ERI.

Some, but not all, clusters of aftershocks following a mainshock earthquake can serve as nuclei for future large earthquakes, as suggested by Empirical, Epidemic Type Aftershock Sequence (ETAS) models. However, it is not currently possible to identify which aftershock clusters might or might not produce future large earthquakes because the variables affecting this phenomenon are not known. Quantified uncertainty bounds associated with ERI's aftershock probability estimate are not available, although ERI acknowledges a "large margin of error".

The Japanese government's probability estimate for a M7.0 earthquake in the vicinity of Tokyo is 70% in 30 years, as calculated by the Headquarters for Earthquake Research Promotion (HERP). This is based on the long-term historical occurrence of events within the estimated magnitude range, rather than extrapolation beyond the range of observed earthquakes. The ERI statement acknowledges that the HERP probability has not changed, since there have been no earthquakes in the Tokyo area of this size since the 2011 Tohoku earthquake.

ERI indicates that an intended objective of releasing the probabilities was to encourage public disaster preparedness related to short-term risk of aftershocks.

Subscribe to CatWatch email alerts and receive CatWatch Catastrophe Reports to your inbox.

Back to CatWatch Reports »