Tsunami, Mega-Earthquakes and Modeling: Tohoku Japan Earthquake 1 Year Later


On this one-year anniversary of the M9 ”Great Tohoku” earthquake and consequential tsunami, we remember the more than 15,000 people whose lives it claimed and reflect on lessons for catastrophe modeling going forward. As one of the most powerful and devastating earthquakes ever recorded, Tohoku-oki struck roughly 40 miles off the east coast of northern Honshu, Japan. The ensuing tsunami triggered a multi-reactor meltdown at the Fukushima power plant, which sent a subsequent shock wave through Japan's infrastructure and economy. Economic damages from the Japan earthquake are estimated to exceed $200 billion USD, and insured losses are estimated to exceed $30 billion USD.

Tsunami Damage - Tohoku Japan Earthquake of March 11, 2011

Tsunami Damage - Tohoku Japan Earthquake of March 11, 2011
Source: Insider Quarterly

Expect the Unexpected

Expect the unexpected. This lesson applies to all recent earthquakes, but is especially relevant after Tohoku-oki. The earthquake's magnitude (M9) exceeded the maximum magnitude (8.3) adopted by the Japanese government (Headquarters for Earthquake Research Promotion) and generally accepted by the scientific community for the fault segment that ruptured. Robust earthquake models for Japan already account for uncertainty in maximum magnitude, but now with this outlier as a benchmark, model updates currently underway capture an even broader range of possibilities.

The extensive tsunami barriers and protections, rationally designed for much shallower flood depths, were overwhelmed by the Tohoku-oki tsunami, which inundated parts of the Japan coast with 25 meters (80 feet) of water. Earthquake damage extended throughout northern Honshu and continued into Yokohama Bay and into Tokyo. Aftershocks included several M7+ earthquakes, and have generated concerns for more earthquakes.

Models Accounting for Uncertainty

After witnessing such "surprises" then, it should be no surprise to the industry that such "known unknowns" will be agents of major damage and loss in future large earthquakes. The challenge is in quantification. Comprehensive scenario planning, coupled with the use of models which conduct a full accounting of uncertainty, are prudent first steps toward diminishing the impact of surprises.

When pondering a possible "worst case," it is notable that the Sanriku coast of Japan - the region hardest hit by Tohoku-oki - does not represent the greatest concentration of earthquake risk in Japan. For example, a re-occurrence of the 1923 Great Kanto earthquake affecting the capital city Tokyo could produce insured losses six times those from Tohoku-oki, or 140 trillion yen.

Models Accounting for Tsunami Losses

Several of the earthquake zones along Japan's eastern coast have demonstrated the potential to produce tsunami, including the Japan Trench, the Sagami subduction zone (near Tokyo), and the Nankai Trough further south. Damage to assets onshore is caused both by the depth (height) of water, as well as its velocity. Future earthquake risk models will account for tsunami-generated losses, especially in areas where the earthquake-induced flood inundation is insured.

Tsunami Wave Amplitude of March 11, 2011

Tsunami Wave Amplitude of March 11, 2011
Source: NOAA Center for Tsunami Research

Mega Earthquakes

The Tohoku-oki earthquake in Japan is the fourth earthquake with a magnitude exceeding 8.5 to have occurred in a span of less than 7 years after more than 40 previous years of “quiescence,” during which not a single earthquake worldwide exceeded M8.5. This observation has initiated a heated scientific discussion on the randomness versus cyclicality of a "global temporal clustering" phenomenon applicable to “giant” earthquakes. Insurance implications of global temporal clustering remain far from being settled, especially considering that a limited subset of earthquake sources can generate an earthquake exceeding M8.5. Ultimately, the debate will broaden our collective understanding of the nature of earthquake mechanics.


More than 1,000 aftershocks have occurred following the main earthquake, with more than 70 shocks at magnitude 6 or greater. The increased frequency of earthquakes following the main event has heightened earthquake awareness and preparation. Among the various published estimates of increased earthquake probabilities for Tokyo, all contain large inherent uncertainties, particularly those based on aftershock occurrence. A probability calculation based on aftershock occurrence in particular is not indicative of medium- and longer-term earthquake risk, nor does it represent the standard of practice for earthquake risk modeling. However, forthcoming updates to Japan earthquake models will include appropriate revisions to time-dependent frequencies, based on peer-reviewed science as soon as it becomes available.

Subduction Zone Plate Boundary Between the Pacific and North American Plates

Subduction Zone Plate Boundary Between the Pacific and North American Plates
Source: U.S. Geological Survey (USGS)

Earthquake Model Development

Each major loss-producing catastrophe increases the sophistication of models and model users. From claims to shaking records, the Japan earthquake and tsunami has created a wealth of data that will inform and improve earthquake model development. From a broader perspective, this event is a reminder that models are only one of many tools to guide decisions that must account for a host of considerations. To holistically assess possible outcomes, risk owners will need to go beyond current capabilities of models.


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

Back to CatWatch Reports »