Total ionospheric electron contents (TEC) were measured by global positioning system (GPS) dual‐frequency receivers developed by the Jet Propulsion Laboratory (JPL). The measurements included P‐code (precise ranging code) and carrier phase data for six GPS satellites during multiple 5‐hour observing sessions. A set of these GPS TEC measurements were mapped from the GPS lines of sight to the line of sight of a Faraday beacon satellite by statistically fitting the TEC data to a simple model of the ionosphere. The mapped GPS TEC values were compared with the Faraday rotation measurements. Because GPS transmitter offsets are different for each satellite and because some GPS receiver offsets were uncalibrated, the sums of the satellite and receiver offsets were estimated simultaneously with the TEC in a least squares procedure. The accuracy of this estimation procedure is evaluated, indicating that the error of the GPS‐determined line of sight TEC can be at or below 1×1016 el/m2. Consequently, the current level of accuracy is comparable to the Faraday rotation technique; however, GPS provides superior sky coverage.

Calculations using a statistical model of water vapor fluctuations yield the effect of the dynamic wet troposphere on radio interferometric measurements. The statistical model arises from two primary assumptions: (1) the spatial structure of refractivity fluctuations can be closely approximated by elementary (Kolmogorov) turbulence theory, and (2) temporal fluctuations are caused by spatial patterns which are moved over a site by the wind. The consequences of these assumptions are outlined for the very long baseline interferometry (VLBI) delay and delay rate observables. For example, at 20 ø elevations in midlatitudes, the wet troposphere induces about 2 cm of delay fluctuation for two-station, 3-hour observations. At 20 ø elevations for 200-s time intervals, water vapor induces approximately 1.5 x 10 -xa s/s of delay rate fluctuation for two-station observations, which corresponds to an Allan standard deviation of 1.2 x 10-t3 s/s. The statistical model suggests that delay rate measurement error is dominated by water vapor fluctuations for most VLBI experimental situations. Water vapor induced VLBI parameter errors and correlations are calculated as a function of the delay observable errors. For example, intercontinental baseline length parameter errors due to water vapor fluctuations are of the order of 2-3 cm. The above physical assumptions also lead to a proposed method for including the water vapor fluctuations in the parameter estimation procedure, which is used to extract baseline and source information from the VLBI observables. the differential phase of an electromagnetic wave be-phase delay rate observables will in turn suffer errors tween two antennas on the earth's surface. Typically due to the fluctuating water vapor, or wet tropothe waves originate from compact extragalactic radio sphere. Typical estimated parameters include basesources, such as quasars. The inferred group delay, line length, baseline orientation, radio source posiwhich is the primary VLBI observable, reflects the tions, and clock offsets [Shapiro, 1976; Sovers et al., distance between the two antennas as well as the 1984]. In most intercontinental VLBI data reduction, angle between the vector connecting the antennas zenith tropospheric delay parameters are also estiand the vector pointing to the radio source [Gold, mated. This technique essentially accounts for a spa-1967; Rogers, 1970; Thomas, 1972]. In astrometric tially and temporally average troposphere, for each and geodetic applications, delays other than the geo-station, for the time over which the zenith parameter metric delay are regarded as errors in the VLBI is estimated. The wet troposphere fluctuations measurement. The presence of atmospheric water around these averages are then the dominant tropovapor along the lines of sight from each antenna to spheric errors, which map to errors in VLBI astrothe source will affect the index of refraction of the metric and geodetic parameters. traversed medium, and will therefore corrupt the geo-When centimeter-level VLBI observable ...

We have used the Very Long Baseline Array (VLBA) at 43, 23 and 15 GHz to measure the solar gravitational deflection of radio waves among four radio sources during an 18-day period in October 2005. Using phase-referenced radio interferometry to fit the measured phase delay to the propagation equation of the parameterized post-Newtonian (PPN) formalism, we have determined the deflection parameter γ = 0.9998 ± 0.0003 (68% confidence level), in agreement with General Relativity. The results come mainly from 43 GHz observations where the refraction effects of the solar corona were negligible beyond 3 degrees from the sun. The purpose of this experiment is three-fold: to improve on the previous results in the gravitational bending experiments near the solar limb; to examine and evaluate the accuracy limits of terrestrial VLBI techniques; and to determine the prospects and outcomes of future experiments. Our conclusion is that a series of improved designed experiments with the VLBA could increase the presented accuracy by at least a factor of 4.With the demonstrated accuracy of the VLBA to measure relative positions of radio sources to 0.01 mas (Fomalont & Kopeikin 2003; Brunthaler et al. 2006), the gravitational bending could be measured potentially with the radio-interferometric technique to a few parts in 100,000, although the coronal refraction when observing within a few degrees of the sun produces large path-length changes. In this paper, we present the results from the measurement of γ performed with the Very Long Baseline Array (VLBA) in October 2005, and we suggest how more accurate measurements of γ can be obtained.

No abstract