Paper presented at the EUREF symposium held in Florence, Italy, 27 - 30 May 2009 Session 6 – Improvement and Densification of ETRS89 1 The new Italian geodetic reference network (RDN): a comparison of solutions using different software packages Maurizio Barbarella, Stefano Gandolfi * , Luciano Ricucci, Antonio Zanutta Dept. DISTART, University of Bologna, Viale Risorgimento 2, 40136 Bologna, Italy ABSTRACT In 2009, the IGMI (Italian Geographic Military Institute) defined a new geodetic reference frame (ETRF2000) based on the new Italian geodetic reference frame RDN (Rete Dinamica Nazionale), a network consisting of about 100 permanent GNSS stations homogeneously distributed throughout the Italian peninsula. A key characteristic of the RDN network is its dynamic nature: the coordinates of the GNSS stations will be computed periodically to take into account the natural changes of the crustal surface. To date, a first computation of the RDN has been implemented during January 2009 by the IGMI using a dataset of 28 Julian Days and the Bernese scientific software. For the validation of the network, other Analysis Centers have been involved in the computation using the same software. One of these Analysis Centers is the Dept. DISTART (University of Bologna) that for this computation has used tree different software packages. This paper presents data- processing results for the same dataset (kindly provided by the IGMI) obtained using Gipsy-Oasis II, Bernese V.5.0 and Gamit-Globk software packages. In order to be aligned with the EUREF guidelines for network densifications an harmonization of the ancillary products and options has been necessary. For the results comparison, some statistical parameters derived by the solutions have been considered. The first comparison has been performed between the Bernese solution of IGMI and the DISTART Dept. one: the differences are at millimeter level. Concerning the comparison of solutions derived by different software packages a very good agreement (at millimeter level) between Bernese and Gamit has been found. Gipsy solution, obtained in PPP and without any ambiguity resolution, presents an slightly higher scattering (but still at millimeter level) than others two. Key words: GPS, network densification * Corresponding author: Prof. Stefano Gandolfi, E- mail address: stefano.gandolfi@unibo.it 1. Introduction In 2008, the IGMI (Istituto Geografico Militare Italiano), the authority for the definition and maintenance of the Italian Geodetic Reference Frame, decided to define a new national reference frame based on a network of permanent GPS stations. This network is called RDN (Rete Dinamica Nazionale). RDN consists of 85 stations homogeneously distributed throughout the Italian peninsula and selected by IGMI (http://87.30.244.175/rdn/rdn.php) from among existing stations installed and maintained by public institutions (such as ASI, INGV, Universities and high schools) and many private companies (Figure 1). To define the reference frame, a dataset of 28 days (from December 23, 2007 to January 19, 2008) was chosen. To frame this network into the EUREF reference frame, computations were performed considering 13 permanent stations of the European Permanent Network (EPN) and particularly (CAGL, GENO, GRAS, GRAZ, IENG, LAMP, MATE, MEDI, NOT1, PADO, SOFI, WTZR, ZIMM) that are located on the Italian peninsula or on the adjacent regions. To officially define the RDN network as a densification of the EUREF reference frame, certain a priori procedures defined by the EUREF TWG (Technical Working Group) are mandatory. In particular, the network computation must be performed by different Analysis Centers and must satisfy certain data processing constraints published on the EUREF web page (http://www.euref-iag.net/). The IGMI has performed the first official computation of the RDN using the Bernese Software Version 5.0 (BSW50; Dach et al., 2007). Two other Analysis Centers (the University of Padua and Polytechnic of Milan) have computed independent solutions of the network based on identical input data and the same scientific software. A third analysis center (a research group of the Dept. DISTART of the University of Bologna) is responsible for processing the same data set using two different scientific software packages: Gamit/Globk Version 10.34 (Herring et al., 2006a, Herring et al., 2006b) and Gipsy-Oasis II Version 5.1 (Zumberge et al., 1997). This research group has also used the Bernese software, following an internal procedure designed to compute a network using additional software systems (Barbarella et al., 2009). The choice to use multiple software pachages has previously been applied in geodetic infrastructures (e.g., Craymer, Piraszewski, 2001; Jivall et al., 2005) and for tectonic or geodynamic purposes (e.g., Dietrich et al., 2001; Even-Tzur et al., 2004; Geirsson et al., 2006; Kierulf et al., 2008, 2009; Simons et al., 1999; Teferle et al., 2008; Zakarevičius et al., 2008, Zanutta et al., 2008, Capra et al., 2008). Although it requires