Analysis of VLBI data with different stochastic models.

Nataliya Zubko, Finnish Geodetic Institute Markku Poutanen, Finnish Geodetic Institute Johannes B\"ohm, Vienna University of Technology Tobias Nilsson, Vienna University of Technology

The VLBI observations are generally analyzed using least-squares method. For the accurate results the functional and stochastic models need to be well defined. In the standard stochastic model the variance- covariance matrix is dependent on only one stochastic parameter, describing by common level of variance. The analysis of observations can be improved by taking into account additional parameters in the stochastic model, such as the station and elevation angle dependent effects. Thus the model becomes reliant on several stochastic properties. A stochastic model, which includes station and elevation angle dependence of observations, has been implemented in VieVS software. We present results of a comparative analysis using traditional and advanced stochastic models. In advanced stochastic model the variance components of VLBI observations were estimated with the simplified MINQUE method.