>> Download full Ph.D. dissertation
>> Download presentation of the Ph.D. defense
>> Latex files
>> Citation: Teke K., (2011). Sub-daily parameter estimation in VLBI data analysis, Geowissenschaftliche Mitteilungen, Heft Nr. 87, Vienna University of Technology, Vienna, Austria. Grafisches Zentrum HTU GmbH, ISSN 1811-8380.
ABSTRACT of the Thesis
Title: “Sub-daily parameter estimation in VLBI data analysis”
The main objective of the work carried out within the scope of this thesis is the contribution to the VLBI2010 project of the IVS by means of developing a parameter estimation module (vie lsm) of Vienna VLBI Software (VieVS) which is capable of estimating accurate sub-daily VLBI geodetic parameters. The vie lsm module is based on the classical Gauss Markoff Least-Squares (LS) adjustment method by using continuous piece-wise linear offset (CPWLO) functions which are estimated at unique epochs, e.g. at integer hours, or at integer fractions or integer multiples of integer hours. The interval for CPWLO modeling of the parameters is usually set to values between one day to five minutes.
To investigate the sub-daily tidal motions of the VLBI antennas during IVS-CONT05, hourly CPWLO TRF coordinates of the antennas were estimated. Although all tidal displacements are computed from state-of-the-art geophysical models and reduced from the observations a priori to the adjustment, the radial amplitudes from the estimated hourly antenna coordinates can reach up to 1 cm (Kokee, HartRAO, Gilcreek, Westford, Svetloe, and Wettzell). To analyze the sub-daily ERP estimates of VieVS during IVS-CONT08, hourly CPWLO ERP were estimated. The Fourier spectra of the hourly VLBI and GPS ERP estimates and the high frequency (short period) ERP models during IVS-CONT08 are in a good agreement at prograde and retrograde 12 hours both for LOD and polar motion. However, at 24 hours prograde polar motion the amplitude from VLBI is larger by about 100 μas than GPS and larger by about 150 μas than HF-ERP models. Additionally, VieVS LOD and polar motion estimates are noisier than from GPS. This may be due to the fact that no relative constraints between the CPWLO ERP estimates in VLBI analysis were introduced. The estimation of hourly source coordinates was rather intended as test study. As long as hourly CPWLO coordinates of two sources are estimated and the remaining sources are fixed to their a priori CRF, the EOP parametrization is not critical for the estimated source coordinates. However, investigations on this issue need to be carried out in the future, e.g. a lot can be learned from correlations between hourly source coordinates and observation geometry.
The second aim of this thesis, which is also very good rest of the CPWLO estimates of troposphere zenith delays and gradients, is the contribution to combination studies in the framework of the Global Geodetic Observing System (GGOS) of the International Association of Geodesy (IAG) by multi-technique comparison of ZTD and troposphere gradients. In the scope of this issue, VLBI VieVS estimates of troposphere zenith total delays (ZTD) and gradients during IVS-CONT08 were compared with those derived from observations with the Global Positioning System (GPS), Doppler Orbitography and Radiopositioning Integrated by Satellite (DORIS), and water vapor radiometers (WVR) co-located with the VLBI radio telescopes. ZTD and gradients estimated by space geodetic techniques are compared to those computed by ray-tracing through the profiles of various Numerical Weather Models (NWM), such as the European Centre for Medium-Range Weather Forecasts (ECMWF) (all sites), the Japan Meteorological Agency (JMA) and Cloud Resolving Storm Simulator (CReSS) (Tsukuba in Japan), and the High Resolution Limited Area Model (HIRLAM) (European sites). The best inter space geodetic technique agreement of ZTD during IVS-CONT08 is found between the combined IVS and the IGS solutions with a mean standard deviation of about 6 mm over all sites, whereas the agreement with numerical weather models is between 6 and 20 mm. The standard deviations are generally larger at low latitude sites because of higher humidity, and the latter is also the reason why the standard deviations are larger at northern hemisphere stations during IVS-CONT08 in comparison to IVS-CONT02 which was observed in October 2002. The assessment of the troposphere gradients from the different techniques is not as clear because of different time intervals, different estimation properties, or different observable. However, the best inter-technique agreement is found between the IVS combined gradients and GPS solutions with standard deviations between 0.2 mm and 0.7 mm. As mentioned before all the comparisons and validation tests on the troposphere products during IVS-CONT08 presented in this thesis provide important information with respect to the planned combination and integration of various observing techniques in the framework of the Global Geodetic Observing System (GGOS) of the International Association of Geodesy (IAG).