Survey Strategies

It is obvious from Figure 1 that the Parkes surveys will take us to a new level of sensitivity over the entire range of HI masses. What the curves do not show is how the strategies of these surveys also affect their results.

Many of the surveys were actually carried out in continuous scanning mode. These surveys can be identified by the -sign in front of the Number of Points in Table 1, which corresponds to the length of the survey strip divided by the beam width. Most of these surveys were done in drift-scan mode, which helps maintain more stable characteristics in the baselines. The Arecibo Dual-Feed survey (in preparation) further improved interference monitoring by using two 21 cm feeds simultaneously, and comparing the output in both polarizations and in the two feeds. These techniques are important for reducing the number of false signals in the surveys.

The Parkes Surveys will have 12 comparison dual-polarization feeds, so that interference monitoring should be excellent. Baseline and standing wave variations may be more problematic if the Southern Sky Survey is done in a driven scanning mode, and this problem may also arise in the Zone of Avoidance survey. In the Arecibo surveys, the Sun produced standing waves between the dish and feed platform that depended sensitively on the sun-dish-feed angles, reducing the survey effectiveness. In continuous scanning modes, however, a model of the standing waves can be generated if they change slowly enough (Briggs et al. 1997).

Spitzak (1996) used a ``step-stare'' mode which achieves much the same effect as the continuous scanning mode. In addition, by ``leap-frogging'' over points on the observing grid, it allowed interference monitoring by requiring that a potential signal not repeat over the short time interval between observations that were separated by large angular distances. Spitzak's survey also is the only deep survey to date that had essentially complete coverage over a large contiguous region. This will also be a feature of the Parkes surveys and is very helpful in determining the luminosity function since it is unnecessary to form complex models of the beam sensitivity for an uncertain displacement of the source from beam center.

Also note that the VLA survey of Weinberg et al. (1991) at first appears comparable to the much earlier Green Bank surveys of Shostak (1977) and Fisher & Tully (1981). The volume limit on this survey was imposed mainly by limitations on bandwidth forced by ``back-end'' correlators and computers. However, Weinberg et al. achieved a much higher spatial resolution than any of the other surveys: the 30' beam size refers to the primary beam--the synthesized beam was . They detected a surprisingly large number of low-mass objects within their small search volume: nine dwarfs, mostly companions to targeted bright galaxies. Large-beam surveys like at Parkes would have trouble separating most of these objects from the larger neighbor. In the multibeam survey it will especially important to achieve Nyquist spacing between beam positions or to tilt the array relative to the scan path in scanning mode to generate sub-beam spacing. This will provide information that should help in resolving source confusion. This is not an area that has been explored by earlier surveys, so I anticipate that a major component of the Parkes survey analysis will be in developing techniques to identify and distinguish confused sources.

© Copyright Astronomical Society of Australia 1997