Introduction

The Italian peninsula is a rather interesting natural laboratory for geophysical investigations. Its tectonic evolution is driven by the interplay of two major plates, the African and Eurasian plates, and possibly by smaller intervening micro plates. The entire area is characterized by a complex tectonic setting where two very different orogens, the Alps and the Apennines, interfere and cause vast areas to deform in a complex way. Compressional regimes are contiguous to extensional regimes along the whole Apennine belt; and to the North, a double vergent thrust belt characterizes the Alpine belt. The whole area is subjected to slow crustal deformations (at the few mm/yr level) (Serpelloni et al., 2005; Devoti et al., 2008; Jenny et al., 2006) originated by the African-Eurasian convergence and modulated by the double subduction of Europe underneath the Adriatic plate and the westward Adriatic plate subduction beneath the Tyrrhenian Sea. The crustal velocity gradients (strain-rates) have been demonstrated to be strongly correlated to the actual seismic activity (Kremer et al., 2002; Bird et al., 2010). The magnitude and kinematics of the strain-rate and consequent stress accumulation on seismogenic structures in Italy is not well known and only recently, thanks to the relatively low cost of GPS surveys, fault-scale mapping of crustal strain-rates have been made available to the scientific discussion and can be tentatively correlated with geological and seismological deformation in an attempt to provide valuable data to help investigating earthquake recurrence and seismic hazard (Caporali et al., 2003, Serpelloni et al. 2005, D’Agostino et al., 2009).

The first attempt to build a nation-wide continuous GPS network was undertaken by the Italian Space Agency (ASI) in the late 1990’s. Since then, it delivers continuous GPS data from about 30 sites and maintains the regional reference frame in strict cooperation with the European reference frame consortium (EUREF). In 2001, the Istituto Nazionale di Oceanografia e Geofisica Sperimentale (INOGS) started installing a local GPS network in the Friuli region (northeastern Italy) to study the deformation pattern of the peri-Alpine thrust. In 2004, the Istituto Nazionale di Geofisica e Vulcanologia (INGV) started the construction of the first national GPS network (RING) dedicated to geodynamic investigation of a wide area (Avallone et al. 2010). At present, the network consists of about 130 stations whose data are continuously transmitted to an archiving center (http://ring.gm.ingv.it) that performs quality checking and data storage. Currently, the data from only 36 of the sites are freely provided on the web site, but there is a strong demand for access to the full archive on a public domain to stimulate the research on this interesting area. Finally, in past years an increasing number of permanent GPS sites were installed by regional administrations and private companies, dedicated mainly to topographic applications and commercial services. These networks, although not conceived to measure long term ground deformations, proved useful in augmenting the backbone RING geodynamic network and are currently archiving their data, making it available for the scientific community. These GPS datasets are currently archived at different INGV archiving centers providing over 400 RINEX files per day for a mean geometric interdistance of about 20 km over the whole country.

The processing of the entire GPS dataset has been carried out by two different analysis centers at INGV (CNT-Bologna and CNT-Roma) using different GPS analysis software (Gamit and Bernese, respectively) and slightly different procedures and models. A rigorous combination of different independent solutions is fundamental in order to cross-validate them, and to produce a final combined velocity field representing the most reliable kinematic representation of the region.

In this work, we perform our analysis following a three-step procedure. In the first step, the GPS raw observations are processed independently using the two software programs following a distributed session approach (Dong et al., 1998), obtaining daily loosely constrained site positions of the different networks. In the second step, the daily solutions are transformed into a common reference frame (ITR2005) to form the position-time series, and then two independent velocity fields are estimated. In the third step, the two velocity fields are rigorously combined in a least-squares sense, thus obtaining the best unbiased estimate of the surface velocity field. All the analysis and combination procedures are performed using the full covariance matrix and following the basic procedures commonly used for the space geodesy reference frame realization and combination (Altamimi et al., 2007).

Because the three steps propagate the full covariance matrix from the original daily solutions, the final combined velocity retains the full information content derived from the observations and from the reference frame datum. Often this type of solution is termed as a ‘rigorous’ geodetic solution or a ‘rigorously’ derived solution (IERS Technical Note 30). As a consequence of the covariance propagation, each site is re-weighted against the entire network so that the final velocity field contains, indeed, the complete covariance matrix.