Australia Telescope National Facility

1. Introduction

The opacity of the atmosphere above the Paul Wild Observatory has been monitored using a 30 GHz water vapour radiometer on loan from Telecom Australia. The radiometer has been in use since September 1991 and, except for about 4 weeks down-time due to technical problems, data have been recorded continuously. In this first report we do not attempt to summarise all data but rather to distill the component most relevant to AT users. In particular, the results presented are for night-time hours of the coldest months. While our overall results show that other times may be usable for 3 mm observations, this block of time is the most favourable. As well as the water vapour data, a limited amount of Compact Array data relating to atmospheric phase stability tests has been analysed; early indications are that the power-law phase versus baseline relationships established at other observatories apply to the relatively low-altitude (205 m above mean sea level) Narrabri site.

2. Results

Opacity data from the 30 GHz radiometer have been scaled to yield 100 GHz values according to the model proposed by Liebe (Rad. Sci., 20 (5), 1985, pp. 1069-1089). At the low opacities of interest to astronomers the model is quite accurate and, if anything, slightly over-estimates the 100 GHz opacity.

Figure A2-1 shows the fraction of time that the 100 GHz night-time atmospheric opacity is less than a given value at Narrabri for the months April - October 1992. Also shown is a representative plot for the proposed NRAO Millimeter Array site of South Baldy, New Mexico (elevation 3200 m). The best months at Narrabri were clearly July and August; the advantage of the much higher South Baldy site is obvious.

Figure A2-2 is an attempt to estimate the number of nights when the atmosphere at Narrabri was stable, at approximately the ±2% 100 GHz transmission level, for a period of 12 hours or more. We have shown the data in histogram form since we suspect that nights with overall atmospheric transmissions greater than 75% form the usable times for aperture synthesis using only conventional (cm-wave style) calibration procedures. Of course, much more time is probably usable with more advanced techniques.

Figure A2-3 is a first attempt at assessing the phase structure of the atmosphere at Narrabri. The plot shows interferometer phase fluctuation as a function of baseline for a six-hour period (1500 - 2100 UT) on July 7, 1992: a night available to us for system test purposes. This night would not have been selected as a particularly good millimetre-wave observing night on the basis of water vapour radiometer data. Furthermore, the array configuration in place at the time did not yield many shorter spacings. Despite this, the plot obtained is rather similar in form to that noted at, for example, the VLA or Nobeyama. The widely-used Kolmogorov model for turbulence seems to fit quite well, with the initial power-law slope in the Narrabri data being ~0.6 and the outer scale length being ~2 km. Much more data is needed but it appears that, for the <1 km baselines dictated by source brightness considerations at 100 GHz, the turbulence behaviour is not anomalous.

3. Future Work

We propose to continue the atmospheric monitoring program for at least two years. The intention is not just to acquire statistics but to use the results to formulate phase and amplitude calibration techniques. The results of this work may have eventual implications for the CA receiver design (e.g. provision of simultaneous 3 mm and cm-wave observing). In the coming year we see a need to make much greater use of interferometer phase measurements, hence our time requests outlined in the main body of this report.

In the next few months we will begin development of a new type of precision water vapour radiometer. This instrument, based on a switched front-end and a digital back-end, will be much more suitable for ATNF site survey purposes than the existing Telecom instrument. The design will also be the prototype for radiometers which could eventually be fitted to the CA antennas, allowing at least coarse on-line phase corrections (based on columnar water vapour measurements) to be made.

1996 Comments

1. Telecom reclaimed the 30 GHz WVR for site testing in Darwin. Owing to some reliability problems with the instrument, little more additional data was obtained.

2. Some ATCA time was allocated for the phase tests; technical problems with the array during the main allocation were troublesome but the phase-baseline plot included in the original report was quite typical. With the array in a very compact configuration on an excellent observing night, the LO phase noise floor was visible at about 0.25 deg. rms per GHz observing frequency (this will be improved with the new mm-wave LO generation and distribution system).

3. The new ATNF radiometer was completed and tested in 1995. See the web pages detailing the phase correction systems. The WVR could indeed be used at Narrabri in the coming winter season (ie while WVR #2 is under construction).

4. I believe that, after I left Narrabri, John Whiteoak arranged a round of systematic phase characterizations.