Seismic moment that accumulates on locked faults during the interseismic period must be released in large seismic events. Measurements of surface velocity from GPS and InSAR can be inverted for locked patches but the inverse problem has the same non-uniqueness and instability as gravity inversion. In this talk we show that moment accumulation rate can be measured from surface velocity without inversion. The accuracy of this measurement relies on the accuracy and spatial coverage of GPS and InSAR data. For example, space geodetic data suggest that significant moment is accumulating along the Creeping Section of the San Andreas Fault but the InSAR measurements must be improved to reduce the uncertainty. This will happen over the next 2 years with data being collected three new InSAR satellites ALOS-2, Sentinel-1A, and Sentinel-1B.

Earthquake and tsunami early warning algorithms that rely on geodetic data for rapid assessment of large sources are becoming widespread. They are, however, not often exercised because GNSS data have noise-levels in the ~2-3cm range and thus are only sensitive to Mw6.5+ events. As a result it is difficult to assess the reliability and performance of these algorithms. I will show an application of the Karhunen-LoÈve expansion to generate stochastic slip distributions of large events with an example application to the Cascadia subduction zone. I'll discuss how to extend the static slip distributions obtained from the K-L expansion method to produce kinematic rupture models and generate synthetic long-period displacement data at the sampling rates of traditional GNSS stations. I'll discuss how to validate the waveforms produced by this method by way of comparison to a displacement based GMPE based on peak ground displacement measurements from GNSS measurements of large earthquakes worldwide. The goal is that this method be used to generate synthetic waveforms for hundreds of events that can be used as diagnostic tests of geodetic based earthquake and tsunami early warning algorithms.