HI-Observations of Obscured Galaxies at Parkes

The redshifts of the uncovered galaxies are obtained by three observational approaches: (a) multifiber spectroscopy with OPTOPUS and MEFOS at the 3.6m telescope of ESO in the densest regions, (b) individual spectroscopy of the brightest galaxies with the 1.9m telescope of the SAAO, and (c) 21cm observations of extended low surface-brightness (LSB) spirals with the 64m Parkes radio telescope. So far we have obtained 1083 new redshifts in our search area.

An extensive discussion of methods and typical results are given in Kraan-Korteweg et al. 1994. The three observing methods are complementary in galaxy populations, characteristic magnitude and diameter range and the depth of volume they probe. The multifiber spectroscopy gives a good description of clusters and dense groups in the ZOA and traces large-scale structures out to recession velocities of 25000 km . The SAAO and HI-observations cover the bright end of the galaxy distribution and provide a more homogenous sampling of galaxies out to 10000 km .

Our HI-observations at Parkes have similar sensitivities as the forthcoming blind HI-survey with the multi-beam (MB) receiver in the ZOA and therefore merit a more detailed description. These observations are vital in recovering an important fraction of the nearby spiral galaxy population which would otherwise be impossible to map. Although we aimed at complete coverage for the brightest galaxies with the optical spectroscopy, a significant fraction of apparent bright galaxies cannot be traced in this manner. In general, this concerns nearby spirals (and also dwarfs) which - seen through a layer of extinction - are extended, very LSB objects.

We have so far observed 345 spiral galaxies without previous redshift estimates at Parkes. Each galaxy was observed in total power mode, a bandwidth of 32 MHz and the 1024 channel autocorrelator. We generally integrated 30 minutes ON / 30 minutes OFF source. This yielded an r.m.s. after Hanning smoothing of typically 4 mJy - hence comparable in sensitivity to the proposed MB ZOA-survey. However, we often detected strong galaxies in as little as ten minutes. On the other hand, we sometimes observed a source for an hour to get sufficient signal to noise. We found the sensitivity during the day to be significantly degraded compared to the level achieved at night due to high-amplitude standing waves in the spectrum which precluded detections even with a large number of very short integrations.

In 1993 we generally centered the recession velocity at 3000 km and reobserved some of the non-detections at a central velocity of 7500 km in a second step. This resulted in a detection rate of over 50%. In 1994 and 1995, we used two IF's and offset 512 channels of each polarization by 22 Mhz. This resulted in a velocity coverage of 0-10000 km with the 32 MHz in one integration. Although the lower frequencies were often badly disturbed by interference around 8300 km (cf., Fig. 2), this increased our detection rate to nearly 80% in 94! Our observations were considerably less effective in the 95-run due to a strong increase in recurring interference, part of which were generated locally by the then ongoing tests of the pulsar equipment - a problem that demands careful investigation of the Galactic pulsar-survey which will be piggy-backed on the MB ZOA-survey!

A few examples of typical HI-spectra, recurring interference as well as detections at low Galactic latitudes are illustrated in Fig. 2.

 
Figure 2: Typical spectra of galaxies in the ZOA obtained with the 64m Parkes radiotelescope. Integration times (between 5 and 60), Galactic coordinates, observed (absorbed) diameters (in arcsec) and magnitudes in the blue B are indicated. Except for panel a, all spectra are baseline-subtracted. Regularly appearing interference of different strengths are found at 1200, 4450, 4700 and 8300 km .

Panel a and b are pointed observations in the Hydra/Antlia extension, c - f are observations in the GA region. Panel a shows a detection in the ON as well as a detection in the adjacent earlier OFF-position in the crowded nearby Hydra/Antlia-filament. Panel b shows another example of a previously unidentified LSB member of this filament. Note the recurring interference at 1200, 4450 and 4700 km . The interference at 1200 km is found in practically all scans (positive or negative). The final shape often resembles real HI-profiles which may be interpreted as erroneous detections. Moreover, detections of weaker real galaxies within that velocity range are impossible.

Panel d shows 2 detections within one beam in the densely populated GA-region, whereas in panel c only the feature at 5100 km results from the detection of a massive GA-galaxy; the stronger signal at 4450 km is due to interference.

The interference at 8300 km (panel f) constitutes a true problem. Unstable in time and strength, it has perturbed about half of the observations in the higher velocity range. Its strength - sometimes over 10 Jy - causes ringing and baseline wiggles and precludes detections of galaxies in a broad velocity range. This makes an analysis of the detection rate statistics and especially the volume completeness function of the MB ZOA-survey extremely hard.

Still, it can be maintained that the HI-observations recover obscured galaxies deeper in the obscuration layer compared to optical spectroscopy; a natural extension into the fully obscured region will be the MB ZOA-survey. The relatively high detection rate of obscured low-latitude galaxies () forecasts quite a success rate for the MB ZOA survey () - with more than one detection per beam in high-density areas (cf., Fig. 2a and d).

© Copyright Astronomical Society of Australia 1997