Three dimensional attenuation and high resolution earthquake location: Applications to the new Madrid seismic zone and Costa Rica seismogenic zone

Part 1: Event archives and continuous waveform data recorded by the Cooperative New Madrid Seismic Network (CNSMN) from 1995 to 2008 are analyzed with waveform cross-correlation techniques to investigate the spatiotemporal distribution of MD < 2.4 earthquakes in the New Madrid Seismic Zone (NMSZ). The resulting clusters are divided into swarm clusters and repeating earthquake clusters depending on interevent duration of time. Most swarm clusters occur near Ridgely, Tennessee. Other swarms and repeating earthquake clusters occur at proposed fault intersections in the crystalline basement or along strong velocity contrasts. The presence of anomalously high pore-fluid pressure is the most likely cause of swarm activity. Repeating earthquake ruptures are interpreted as reactivation of small asperities.;Part 2: A three-dimensional, high-resolution P-wave seismic attenuation model (QP) for NMSZ is determined from P-wave path attenuation (t*) values of MD < 3.9 earthquakes recorded at 89 seismometers of the CNMSN and 40 seismometers of the Portable Array for Numerical Data Acquisition (PANDA) deployment. The amplitude spectra of all the earthquakes are simultaneously inverted for source, path and site parameters. The t* values are inverted for Qp using local earthquake tomography (LET) methods and a known 3D P-wave velocity model for the region. The four major seismicity arms of the NMSZ exhibit lower Q P values than the surrounding crust. The larger QP anomalies coincide with previously reported high swarm activity attributed to possibly fluid rich fractures along the southeast extension of the Reelfoot fault.;Part 3: We use the spectra of 210 earthquakes recorded by 35 seismometers to image the attenuation structure of the seismogenic zone below Nicoya Peninsula, Costa Rica. The amplitude spectra of the earthquakes are used to estimate t* using common spectrum method. An attenuation map is then obtained using LET using a previously constrained velocity model and earthquake locations. We use the final high-resolution attenuation results from this study to investigate the relationship among attenuation heterogeneity, interseismic coupling and genesis of different classes of earthquakes. Large earthquakes are generated at fully coupled patches of the megathrust characterized by low QP values. Low frequency earthquakes tend to occur in low QP values.