Investigations into a Dynamic Geocentric Datum

Abstract

A "dynamic" geodetic datum is a datum that consists of coordinates and velocities of control points, which can provide users with the realistic positions of sites within the network that reflects the result of local or regional motion induced by a number of causes. A new ITRF-linked scheme is proposed and discussed in order to process available geodetic data in real-time to establish a dynamic ITRF-like datum. The authors suggest providing a variety of such datums for different applications. The noise properties of the geodetic coordinate time series are crucial to reliably estimating positions, velocities, and their stability and uncertainties. A 'db3' wavelet transform is used to separate the noise of the velocity series and the autocorrelation. The histogram distribution and spectrum features of the residuals indicate that the outliers impact the de-noising effect. A Kalman filter-based model with the outliers being detected and removed online is investigated to obtain the "purified" velocity series for ITRF modeling. For a local scale application, interpolation methods play an important role in improving the accuracy of the datum. Six interpolation methods are investigated using a simulated velocity data set. The "local patch" model is necessary for a local datum in order to handle unexpected events. © Springer-Verlag Berlin Heidelberg 2009.