It has been known since pulsars were discovered that the interstellar medium (ISM) can cause angular broadening of radio sources. However, the amount of scattering along typical lines of sight to extra-galactic sources is not large enough to cause any loss of visibility, even on intercontinental baselines, unless the frequency is below about 100 MHz. Recent studies have shown that interstellar scattering (ISS) can have considerably stronger effects; this paper addresses their relevance to VLBI.

We limit the spectral index of density fluctuations in the interstellar plasma by comparing maser sizes, due to small-scale fluctuations, with wander of masers from constant velocity, possibly due to refraction by large-scale fluctuations. The wander of the masers about constantvelocity motion limits power-law spectra of the fluctuations to indices less than 4.0.

Results obtained on the Low Frequency Variability (LFV) phenomenon, by means of combined multifrequency observations of 50 sources, on a period of more than ten years on a frequency grid of 0.4, 2.3, 4.8, 8.0, and 14.4 GHz and two epoch VLBI observations at 18 cm can be summarized as follows:

1. 1. 15–20% of variables appear to have variations consisting either of quasi-simultaneous outbursts at all frequencies or of bursts which drift to lower frequencies with time and decreasing amplitude. In our sample, we find five good cases: 3C 120, 0605-085, 1510-089, 3C 345, BL Lac. Three of these are famous superluminals; the other two show significant structural changes between our 18 cm VLBI measurements. The corresponding expansion rate for these five sources is in agreement with the γ's derived from LFV with the usual causality arguments. For the sources of this class, the observations are therefore in agreement with models that explain the phenomenon of the variability as synchrotron emission of relativistic electrons beamed in a direction close to the line of sight.

2. 2. 35% of variables show only low frequency (<1 GHz) variability and little or no intermediate high frequency variations. In DA 406, prototype of the category, no superluminal motions have been observed, even if the resolution of our VLBI observations should allow the detection of the structural change expected on the basis of intrinsic LFV. In this case we do not find direct evidence of relativistic motions associated with the LFV and the process is most easily explained if the variations are extrinsic (propagation effects through the interstellar medium as the slow refractive scintillation).

3. 3. The remaining 40–45% of variables show uncorrelated high (<5 GHz) and low frequency variability with a minimum of activity at the intermediate frequencies. The explanation of the phenomenon is less clear. It could be attributed to intrinsic (superluminal) variations at high frequencies, coexisting with unrelated processes at low frequencies.

Ray tracing numerical simulations of extreme scattering events result in light curves that display many of the qualitative features observed in such events. The apparent images of an affected source show strong and variable distortions, which could be readily discerned with VLBI observations.

Dally flux density measurements of 36 extragalactic radio sources over a seven year period, obtained by the Green Bank interferometer, reveal several unusual minima in the light curves that do not follow typical source variations (Fiedler et al. 1987). The most significant departure from typical source variability occurred at both frequencies in the quasar 0954+658 between 1980.95 and 1981.3. Refractive focussing by small scale inhomogeneities in an ionized structure in the interstellar medium appears to be the most likely explanation.

We report progress on a study of interstellar scattering in the direction of the constellation Cygnus. A high quality, 1663 MHz VLBI image has been obtained of the radio source 2013+370, whose structure at this frequency is dominated by scattering. The visibility function indicates that the spectral index of the interstellar density irregularities is between 3.7 and 4.0, with the higher value being somewhat more likely. Another project consists of multifrequency observations of eight sources in this region. Pronounced changes in the scattering are observed for sources separated by only a few degrees.

MERLIN observations at 0.408 GHz of the December 1985 outburst of Cyg X-3 have enabled us to characterise the angular broadening, caused by interstellar scattering, very accurately. The scattering along the line of sight towards Cyg X-3 is exceeded by only two other lines of sight through the Galaxy: towards the Galactic centre (see e.g. Lo et al. 1985. Nature, 315, 124) and towards NGC 6334 (Rodriguez et al. 1982. Astrophys. J., 225, 103); as a result quite short baselines are needed to study the scattering disk at 73 cm wavelength (0.408 GHz). The projected MERLIN baselines for our observations cover the range from a few km (on which the source was unresolved and had a correlated flux density ∼3.6 Jy) to ∼130 km (on which the source was totally resolved i.e. a correlated flux density <0.030 Jy). The basic result from these observations is that over this range of baselines the scattering appears to be purely diffractive in character. The scattering disk is, to quite a good approximation, a circular gaussian and shows no evidence of fine scale substructure. Model fitting to the visibility amplitudes, assuming circular symmetry, yields θ0.408GHz = 2.85±0.05 arcsec (FWHM).

Electron irregularities in the solar corona scatter the radiation from radio sources. During more than 30 years of coronal electron irregularity investigations, a very complex temporal and spatial behavior of the solar corona have been observed. In particular, the scattering dependence on distance from the Sun varies with both the phase of the cycle and the different solar maxima. The influence of scattering on the highest resolution observations in radio astronomy, i.e., those made with VLBI, is not negligible. It should be taken into account in planning of both the global VLBI as well as the future orbital VLBI stations such as the QUASAT and RADIOASTRON missions. The “normal” scattering dependence on the distance from the Sun is close to the inverse square law, but substantial deviations from this dependence have been found.

The determination of a specific catalogue or ephemeris reference frame is a highly over-determined problem, depending on the particular selection of which coordinates, which objects and at what time(s) the determination is made. The consistency which various determinations exhibit is dependent upon the accuracy of the catalogue or ephemeris itself. This paper discusses the accuracies of the three most prominent celestial reference frames: stellar catalogues, the lunar and planetary ephemerides and the radio source catalogues.

The FK4 stellar catalogue contains known systematic errors amounting to a few tenths of an arcsecond; the FK5 will yield nearly an order of magnitude improvement; HIPPARCOS and Space Telescope expect by the mid 1990's optical interferometry should approach within a couple of years, tens of micro(!)arcseconds after a couple of decades. Present-day lunar and planetary ephemerides have accuracies at the level of for the moon and inner four planets; for the outer planets. Further observational data will permit continued improvement. Radio source catalogues now show internal consistency of

An almost inertial celestial reference frame based upon extragalactic sources is in the process of being established. This reference frame is to be global with a minimum density of one source/100 square degrees. The source positions will be based upon radio observations and will allow optical reference frames to be related to this frame at the 0.03 arc second level.

A global array of 20 radio observatories was used to measure the three-dimensional position and velocity of the two meteorological balloons that were injected into the equatorial region of the Venus atmosphere by the VEGA spacecraft.

This project would be the next step in our ability to detect very low frequency (VLF) gravitational waves and the first committed spaceborne designed experiment. Present Deep Space spacecraft tracking experiments are severely limited in their detection capability. It is proposed to construct a spaceborne multi-arm microwave interferometer using current elements of design applicable for the detection of VLF gravitational waves. The elements are outlined with particular emphasis placed on the utilization of small inexpensive get away special (GAS) modules currently under development at JPL for launch in the 1990's.

Atmospheric fluctuations and hydrogen maser clock instabilities limit the coherence time (Tcoh) and hence the signal-to-noise ratio of VLBI data , to a few minutes at GHz frequencies. However, Tcoh can be increased up to many hours if a phase reference is established throughout the experiment by using a strong source observed alternatively with the program source. Weak celestial radio sources with flux densities in the milliJansky range at centimeter wavelengths, such as radio stars, fast pulsars and faint components of extragalactic sources, can then be reliably detected by this VLBI technique, simultaneously providing accurate astrometry relative to the reference source.

VLBI currently produces the most accurate positions of celestial objects. From 1979 to 1987, 114 extragalactic radio sources have been observed with dual-frequency Mark III VLBI as part of the NASA Crustal Dynamics Project and the NGS POLARIS/IRIS program. The formal statistical errors of conventional celestial coordinates are as small as 0.3 milliarcseconds. The fundamental quantity measured by VLBI is the arc length between radio sources. Thus, we suggest that VLBI be used to establish a coordinate reference frame based solely on radio positions, and that this system not necessarily be coupled to right ascension and declination.

Radio positions of eight stellar systems have been derived with accuracies of 3 to 300 milliarcseconds from MkIII VLBI observations conducted with multi-station arrays. The best accuracy was obtained during a relatively strong outburst of the RS CVn system HR 5110. The epoch J2000.0 positions are obtained in the Jet Propulsion Laboratory VLBI reference frame of extragalactic radio sources.

Optical positions of eight radio stars observed with the Photoelectric Transit Instrument (PTI) at Torino and with the Danjon Astrolabe (AST) at Cagliari are presented here. The values of observed right ascensions have been computed at the Epoch J2000.0.

Intercontinental radio interferometric measurements between NASA's Deep Space Network stations yield a catalog of positions of 106 extragalactic radio sources, uniformly distributed between −45° and +85° declination. Almost all of the source positions have formal uncertainties between 0.5 and 3 milliarcseconds. Estimates of three independent sets of 32 source coordinates at average epochs 1980.0, 1983.2, and 1984.5 assess the level of stability of the reference frame over a five-year period. Comparisons of the three sets of source coordinates show a number of cases of > 2σ differences, particularly betwen the 1984–85 and 1981–83 periods. No sources, however, show 2σ shifts among both pairs of epochs. Upper limits of the order of 1 mas/yr can be placed on the time rates of change of the 32 source coordinates.

Future deep space missions will benefit from submilliarcsecond angular navigation accuracy (< 25 km projected distance at Neptune). Conventional Very Long Baseline Interferometry (VLBI) measurement strategies entail observations of a radio source (RS) and spacecraft pair. To increase the 3–5 milliarcsecond accuracies currently attainable, instrumental improvements could be coupled with either of two methods: 1) Use improved calibration measurements with the single RS as the spacecraft reference, or 2) use a reference frame established by a local network of RS's near the spacecraft to reduce the contribution of sky-dependent error sources. The demanding baseline vector (earth orientation) and tropospheric calibrations needed to realize submilliarcsecond accuracy with the single-source strategy would not be necessary with the multi-source strategy. The two strategies are compared via a sample covariance analysis. In this analysis, the single-source method yields 1.5 milliarcseconds for the navigation accuracy, while the multi-source method yields 0.5 milliarcseconds.

The HIPPARCOS satellite will produce positions, motions and parallaxes of celestial objects with previously unattained accuracy. This HIPPARCOS Instrumental System, however, will have an unknown solid body rotation with respect to an inertial reference frame. One aspect of our program of astrometric observations with the Hubble Space Telescope is to determine the rotation of the HIPPARCOS reference frame with respect to an extragalactic reference system.

The Crustal Dynamics Project has been developing, deploying, and operating very-long-baseline interferometry (VLBI) systems and satellite laser ranging (SLR) systems for highly accurate geodetic measurements of global plate motion, plate stability, regional crustal deformation, and earth rotation/polar motion. Over the past 10 years, the measurement accuracies of these systems have been improved by a factor of 10 to the cm level. Plans are to continue these developments to reach mm level accuracies. The present deployment of the VLBI systems is primarily in the Northern Hemisphere. This network has produced measurements of the relative plate motion between the North American, Eurasian, and Pacific plates; the stability of the same plates; and the regional deformation at the North American/Pacific plate boundary in California and Alaska.