Long-term remote single-dish observations of blazar radio variability.

This research has demonstrated that a small, remotely operated radio telescope can perform well enough to monitor blazar radio variability over periods of months to years. Such observations are not possible using premier telescope facilities, given observing time demands, and they enable scintillation effects intrinsic to the source to be disentangled from scintillation due to scattering of radio waves by the interstellar medium. This exercise provides insight into the nature of the source, and also provides a probe of turbulence in the interstellar medium. The University of Tasmania's 30 m antenna near Ceduna in South Australia was converted to a radio telescope facility in 1997 from its former use as an Earth station. The COntinuous Single dish Monitoring of Intraday variables at Ceduna (COSMIC) campaign started in March 2003, and extended to early 2005. It observed a number of blazars, with the telescope remotely operated from Tasmania. The blazars were divided into groups lying south and north of the zenith at Ceduna, with each group served by a calibrator source and observed in turn for periods of 10-15 days. A source scanning strategy was developed, and semi-automatic software procedures were written to process raw data into calibrated flux density data sets, corrected for gain-elevation and pointing, and subject to quality control tests. The consistency in calibrator observations over the ~2 year period shows that a 30 m antenna can carry out long term monitoring of blazars with strengths vîvá‚â• vîvÖv¶1 Jy to the accuracy needed to identify variability on time scales of days, and better performance is expected in future campaigns The antenna's 1/f noise is ~1% of the total flux density, and is likely due to electronic gain fluctuations. It is about 2¬¨Œ© times greater than thermal noise at the integration times relevant to the Ceduna flux density measurements. COSMIC campaign data contain vîvᬱ0.15 Jy systematic flux density fluctuations, that have a thermal origin. These fluctuations were initially believed to be genuine variability, and are most evident on diurnal time scales. The raw data processing exercise cannot be adjusted to remove the fluctuations for the blazars of interest to this research, PKS B1622-253 and PKS B1519-273, but the genuine variability in these two blazars occurs on time scales of ~1-10 days. A method of filtering and correcting the flux density data was developed, the strategy being to smooth through the diurnal systematic effects, remove longer term flux density trends, correct for systematic effects on weekly and seasonal time scales, and hence isolate the genuine variability. A suite of variability analysis tools appropriate for Ceduna data was developed, using the scintle peak-to-peak period, Tperiod , to define the characteristic variability time scale. Values of T0.5 or T1/e can also be estimated, enabling examination of decorrelation timescales, but with caveats due to the peculiarities of the Ceduna data sets, whose data gaps and other characteristics provide challenges to an analysis of variability on a time scale of days. - iv - Tperiod values are determined for each 10-15 day observing period by spectral analysis, using a power spectral density function obtained as the Fourier transform of a discrete autocorrelation function. Empirical scintle counting and data folding exercises cross-check the Tperiod values. Scintle periods are well modelled as Gaussian distributions that are similar for the two blazars, since both sources are large enough to band-limited the scintillation process in similar ways. The statistical properties of the scintle periods provide empirical error bars estimates for the Tperiod values. Also, the 95% confidence interval error bars for Tperiod values calculated from a typical set of scintles are comparable to the 2vîvÖ‚â• vîvÖv¶ 10% upper limit of the stochasticity in Tperiod values that Monte Carlo modelling predicted would enable Tperiod to be computed with fair accuracy. For both PKS B1622-253 and PKS B1519-273, the Tperiod values computed for each observing period over the COSMIC campaign exhibit clear annual cycles, which unequivocally proves that in both cases the observed scintillation is primarily due to scattering by the interstellar medium. Multi-frequency observations of PKS B1519-273 have shown that its scintillation is associated with the weak scattering regime at the 6.7 GHz Ceduna observing frequency, and this is also believed to be the case for PKS B1622-253. The annual cycles in the variability time scales (i.e. Tperiod values) of the two blazars are well fitted by the standard model of interstellar scintillation. Tperiod values for PKS B1622-253 and PKS B1519-273 range from about 2 ‚Äö- 10 days and about 1-5 days respectively. The strength of PKS B1519-273 fell below ~¬¨Œ© Jy in mid-2004, precluding accurate determination of Tperiod values in the final months of the COSMIC project. For both sources, the best annual cycle model fit is for highly anisotropic scintles and large velocity offsets of the scattering screen with respect to the Local Standard of Rest. This is unsurprising, since scintillation on time scales of days is associated with distant scattering screens, typically hundreds of parsecs from Earth, which are often in motion with respect to the LSR. The annual cycle model fits to the PKS B1622-253 and PKS B1519-273 Tperiod values have reduced chi-square values of 2.12 and 0.83 respectively, confirming that the empirically determined error bar estimates are appropriate, and that the annual cycle model credibly describes the variation in the variability time scales of the two blazars. The variability characteristics of PKS B1519-273, and the annual cycle in its variability time, agree well with previous analyses of this source based on more limited data, but data recorded by much better telescopes. This agreement confirms the success of the COSMIC project. An annual cycle in the variability time scale of PKS B1622-253 has not previously been observed. The main follow-on research tasks are to study the implications of the variability characteristics of both PKS B1622-253 and PKS B1519-273, with consideration of anisotropy; eliminate the problem of systematic fluctuations; and examine the other blazars monitored in the COSMIC campaign.