Constraining Paleoearthquake Slip Distributions with Coral Microatolls

Abstract:

Key to understanding the threat posed by large megathrust earthquakes is identifying where the potential for these destructive events exists. Studying extended sequences of earthquakes, Slip Deficit and Stress Evolution modelling techniques may hold the key to locating areas of concern. However, as well as using recent instrumentally constrained slip distributions they require the production of high resolution source models for pre-instrumental events.

One place we can attempt this longer term modelling is along the Sunda Trench with its record of large megathrust earthquakes dating back centuries. Coral microatolls populating the intertidal areas of the Sumatran Forearc act as long-term geodetic recorders of tectonic activity. Repeated cycles of stress accumulation and release alter relative sea levels around these islands. Growth of corals, controlled by the level of the lowest tide, exploit interseismic rises in sea level. In turn, they experience die-offs when coseismic drops in sea level lead to subaerially exposure. Examination of coral stratigraphy reveals a history of displacements from which information of past earthquakes can be inferred.

We have developed a Genetic Algorithm Slip Estimator (GASE) to rapidly produce high resolution slip distributions from coral displacement data. GASE recombines information held in populations of randomly generated slip distributions, to create superior models, satisfying observed displacements. Non-unique solutions require multiple iterations of the algorithm, producing a suite of models from which an ensemble slip distribution is drawn. Systematic testing of the algorithm demonstrates its ability to reliably estimate both known synthetic and instrumentally constrained slip distributions based on surface displacements. We will present high-resolution source models satisfying published displacement data for a number recent and paleoearthquakes along the Sunda trench, including the great 1797 and 1833 events.