KALREF - A Kalman filter approach to the International Terrestrial Reference Frame realization

To monitor global geophysical changes with realistic uncertainties using millimeter-precision geodesy, it is essential to define, realize and maintain the International Terrestrial Reference Frame (ITRF) consistently and accurately. Precise determinations of geocentric site positions and motions, satellite orbits, geocenter motion, Earth orientation and its variations, mean sea level rise, and polar ice mass changes at various time scales all depend critically on the accuracy and stability of the ITRF. Currently, the ITRF is a secular frame based on a linear model and consisting of mean epoch positions and velocities for a global set of stations. With higher accuracy and density geodetic data increasingly used for near real time geophysical change determinations, a research version of Terrestrial Reference Frame (TRF) realized by station time series is needed, which combines strengths of different geodetic techniques and unifies the different technique time series in the single frame. Here, we conceptually define an experimental reference frame with its origin at the nearly instantaneous center-of-mass (CM) of the total Earth system, by specifying the frame and combining different technique data weekly (daily for Earth orientation parameters). For co-located sites, available local ties are applied only once; but site motions are usually constrained to be the same. A Kalman filter and smoother algorithm has been developed and coupled to the ITRF/CATREF software to solve for geocentric coordinate time series with a model of secular, periodical and stochastic motion components. Results using linear and linear plus sinusoidal motion models without stochastic components compare very favorably with the ITRF2005 solution. Filtering strategies and time series results using ITRF2005 and ITRF2008 input data will be presented.