Don Percival         Principal Mathematician         EIS Department         Applied Physics Laboratory         University of Washington	Current statistical models and analysis are not in common use in the precise time and time interval (PTTI) community. Important methodological advances in the past twenty years include: (1) the formation of fractionally differenced (FD) processes as tractable models for processes with stationary increments (these are superior analytically to the power law models in common use today); (2) the development of multitaper spectral analysis, which offers stable estimates of the power spectrum with low bias; and (3) the development of wavelet variance analysis, which offers important generalizations of the well-known Allan variance for characterizing clock noise. This project begins with the techniques of spectral analysis, wavelet analysis and maximum likelihood estimation as applied to FD processes. As needed to handle particular noise types, we are investigating more complicated models (e.g., composite FD processes and time-varying FD processes) and additional techniques that will serve to complement or extend our suite of analysis techniques (e.g., to handle various types of deterministic trends). Our emphasis is on techniques and models that are now commonly accepted in the statistical community, but that have yet to be commonly used by clock analysts. With our thorough documentation of our analyses of the various clock noises, we intend offering these as examples of how to bridge the gap between modern statistical theory and practical applications.