Densification of IGS/EPN by local permanent networks: sensitivity of results with respect to the adjustment choices

Local GNSS permanent networks materialize the global reference frame through their estimation and distribution of coordinates and velocities. These estimates are provided by the classical network adjustment process, including several nearby reference IGS/EPN permanent stations as fiducial constraints. GNSS data modelling is still a topic of research and the final results can be very sensitive to the processing choices: during the last years, IERS conventions as well as IGS and EPN guidelines have been periodically updated in order to reflect the state of the art. This work aims at evaluating the sensitivity of the adjustment results with respect to the processing choices. To perform the tests, two local permanent networks, characterized by different geometries and geographic locations are used; the first network is in Lombardia, a region in the northern, pre-alpine Italy; the second is in Puglia, in the southern, Mediterranean, Italy: respectively, 12 and 6 months of data are analysed. Different processing choices are applied and the relevant results are compared: firstly, different constraining weights for the fiducial stations are tested, secondly, the estimation of ZTD’s alone is compared with the estimation including horizontal gradients, and finally the ocean loading effects are analyzed. In the adjustment of a local network, the normal practice is the inclusion and constraining of some IGS stations; we have studied the adoption of a national zero order network as a link between the global and the local networks: the new zero order network of Istituto Geografico Militare (IGM, the Italian Cartographic Institute) is used for an experimental test: in this case, only one month of data is available and is analysed. At the end, the differences between IGS05 and ITRF2005 coordinates are analysed.