The study of intraplate earthquakes is fundamental for the understanding of the physics of faulting, seismic hazard assessment, and nuclear monitoring, but large to moderate well recorded intraplate earthquakes are scarce. I use the best recorded earthquake in Eastern North America (ENA)---the Mw 5.0 20 April 2002, Au Sable Forks, NY, earthquake and its aftershock sequence to investigate wave propagation and earthquake source properties in ENA.;The Au Sable Forks epicenter is located near the boundary of two distinct geological provinces Appalachian (New England) and Grenville (New York). Existing regional one-dimensional (1D) crustal models were derived from seismic surveys or from sparse ground-motions recordings from regional moderate earthquakes. I obtain improved 1D crustal models for these two provinces by forward modeling, for the first time, multi-path high-quality ground-motions of a moderate earthquake in ENA. Using Au Sable Forks earthquake records at 16 stations (epicentral distances < 400 km) at intermediate frequencies (<1 Hz), I generate synthetic seismograms using the frequency-wave number method. The new models improve the fit of synthetics to data at all 6 stations in the Grenville province and at 5 of the 10 stations in the Appalachian province. I identify complex wave paths along the boundary between the provinces, and 3% azimuthal anisotropy in the Appalachian crust.;It is unknown how much earthquake source properties depend on the tectonic setting in which the earthquakes occur. Debate exists regarding the invariance of stress drop with earthquake size in ENA, and whether earthquakes in intraplate regions have higher stress drops than those in more tectonically active regions. I estimate source parameters for 22 earthquakes (M1-M5) of the Au Sable Forks sequence, using two alternative methods: a direct wave method (Empirical Green's Function) and a coda wave method (Coda Ratio) applied for the first time to small magnitude earthquakes. Both methods consistently show high stress drops, radiated seismic energy and apparent stress. Results are significantly larger than estimates for interplate earthquakes, suggesting stronger faults in ENA. I find constant stress drop up to the bandwidth resolution limit (80 Hz) of the study, and no evidence of stress drop breakdown. |