Accuracy and precision vs software and different conditions using Italian GPS fiducial network (IGFN) data

As well known standard deviation estimation of GPS baselines (and components) by dedicated software are usually lower than the "real" accuracy. Assuming as total errors the addition of stochastic and systematic components, GPS software solution for the related incertitude takes into account mainly the stochastic component. Systematical errors are not estimable by observable but can be reduced only by models (or by calibration in the case of measure instruments). For these reasons, accuracy and precision are two different entities and sometime using GPS it is possible to obtain good results in terms of RMS, but not accurate ones. The baseline accuracy is dependent from many factors such as: baseline length, constellation, windows time observation, ionospheric and tropospheric models, etc. All software performs data processing using models and algorithms and, despite of the most of cases models are the same, the algorithm that applies the model could be different (different numerical methods for computation etc…). Not only, frequently the data processing philosophy used by different software is different and these different characteristics are probably derived by different initial requirements. Therefore, if we decide to use GPS for monitoring and particularly for monitoring very small movements the correct approach consists in choosing a software and performs the same data processing procedure for all the surveys. Using that approach all models and all “problems” are evaluated in the same modality and all the results have to be considered “coherent”. Others application do not need these accuracies and if we are interested to decimeter level accuracy these problem are not so evident. Actually, the international scientific community disposes lots of GPS data processing scientific software and each one is characterized by different peculiarity. The utilization of this software is frequently related to the geography and if GIPSY-OASIS II is very common in USA, BERNESE Software plays a main rule in Europe. Another software that recently has enlarged the number of user both in Europe than in United States is GAMIT but other software as PAGES and Microcosmos are also available. This work has to be considered as preliminary and only with an enlargement of samples and cases can be more significant. In this paper the author start with a comparison of two software in different conditions. Naturally the number of parameter that contribute to the solution variability is very large and in this paper only few of these are evaluated. Particularly stability on solution for different windows time observation and for different baseline length have been investigated.