Assessment of GPS Derived Velocity Solutions in the Mediterranean Area

The GPS data collected from permanent receivers are now widely used to compute time series of site coordinates and to infere the kinetic behaviour of tectonically active regions. The daily (or weekly) coordinate solutions are often considered as high level products obtained by a complex propagation of the raw GPS observations into the coordinate space. A number of highly sofisticated GPS analysis tools have been developed in the past years and are now available to a great deal of applications. Nevertheless, differences in the setup approach, misconceptions or real software bugs could lead to hidden errors in the final geodetic product. The apparently simple-to-use softwares could therefore give the impression of easy-to-obtain results and often geophysical interpretations relay on data that have not been cross validated. This work will test the contribution to the error budget of long lasting GPS time series (1995-2004), computed by different analysis groups and different softwares (Gipsy, MicroCosm and Gamit) in a regional network context. The current work focuses on time series of permanent GPS solutions located in the Central Mediterranean area, already published or that have been used for regional tectonic studies. Site velocities has been derived from these time series with a standardized approach and using the complete covariance matrices. The different velocity fields are then compared in a homogeneous reference frame. Eventual differences are purely due to the raw data analysis procedures and should never exceed the given standard error. The site noise content derived from the coordinate time series will also be discussed and compared in order to possibly isolate the error sources. Since the expected velocities in the central mediterranean area are only on the order of a few mm/y, there is a strong demand on the stability of the coordinate and covariance time series. If the repeatability among different data-analysis approaches is not at the level of the declared standard error, the geophysical interpretation can be, at high probability, seriously distorted by impredictible (human) factors.