Better understanding earthquake riskBy Feng Wang | December 29, 2016
Trembling with anticipation of the "Big One," the hypothetical M8.0+ earthquake expected to occur along the San Andreas Fault, California is no stranger to earthquakes. The state is sandwiched between two tectonic plates: the Pacific and North American plates, which shift northwestward and southward, respectively. Approximately 500 earthquakes of the many more California experiences each year are large enough to be felt.
The probability that an earthquake of a specified magnitude will occur on a fault or within a given area can be calculated by projecting earthquake rates based on earthquake history and fault slip rates.
The Working Group on California Earthquake Probabilities (WGCEP) has long been in the vanguard in adapting creative ways of using various earthquake-related information and physical modeling for such statewide earthquake hazard analysis. In 2008, WGCEP published a Uniform California Earthquake Rupture Forecast (UCERF2).
UCERF3, which includes an update of input data and new methodology, followed in 2014. As the UCERF3 documentation notes, the WGCEP's progression of models "reflects increasingly accurate, detailed, and sophisticated representations of a particularly complex natural system.
So what's new in the UCERF3 model? In general, the results from the new study confirm previous findings, but because of improvements to the model there are also some significant changes. UCERF3 uses physical modeling more than its predecessor UCERF2, which relied more heavily on expert opinion. It also considers much smaller rupture units to delineate faults and defines complex multi-segment rupture scenarios where earthquakes are no longer confined to separate, individual faults but can rupture multiple faults simultaneously on occasion. This reflects what seismologists have been observing from earthquakes around the world concerning the complexity of dependent fault interactions and stress triggering at depth for faults where, at the surface, these factors appear to be independent.
Because of the inclusion of multi-fault ruptures, UCERF3 estimates a rate of seismicity about 30% lower for intermediate earthquakes with a magnitude of about 6.7 and an increase of about 20% in the rates of M8.0 and larger temblors compared to the rates published in UCERF2. In addition, UCERF3 now considers more than 250,000 different fault-based earthquakes, while UCERF2 studied only about 10,000 earthquakes, and the previous model far fewer.
How is AIR implementing UCERF3?
In response to the new findings published in UCERF3, building codes, earthquake insurance products, emergency plans, and other risk mitigation efforts will all be updated. UCERF3 has also been a key source of data for the update to the AIR Earthquake Model for the United States, which is scheduled for release next summer. Specifically, UCERF3 introduced more than 1.6 million rupture scenarios on faults for which many have estimated rates lower than 1 in one million years and some lower than 1 in one billion years.
AIR seismologists created and optimized a 100,000-year catalog from this data set using an optimization procedure that produces consistent hazard curves at a number of key locations in California from the full UCERF3 rupture scenarios. The key locations were identified based on population centers, their vicinity to major faults, and their distribution across the state.
For inclusion in the 100k catalog, UCERF3 scenarios were categorized based on their rates of occurrences and the optimization procedures were designed to properly sample the very low rate scenarios.
The soundness of an earthquake model and its ability to produce reliable results rests on the strength and scientific rigor of each of its components. UCERF3 contributes directly to the AIR model's comprehensive and up-to-date view of seismicity. A robust, probabilistic view of the risk enables companies to prepare for the next large-scale event—wherever and whenever it occurs, and whether or not it is indeed the "Big One."
Feng Wang is a Senior Scientist in the Research Department at AIR.
This article was produced by AIR Worldwide and first published in AIR's In Focus blog.