Peter Hall and Jared Thompson
14 January, 1996 (preliminary).
Introduction
We have examined cloud cover data recorded at the Moree weather station for the decade 1985-95. Exact details of daily observation times etc. are available on request but, in general, the results are compiled from 3-hourly visual observations of total cloud. Moree is 100 km north of Narrabri but both centres share the inland continental climate typical of the western districts of NSW. Both centres use common aeronautical weather data and we are confident that the Moree results are representative of the Narrabri situation.
Results
The attached plots show the proportion of time for which the sky is clear, i.e. no more than 1 Oktas (1/8) total cloud. Since we are interested in the amount of time that synthesis observations are possible, we examine the number of unbroken 12-hour clear windows for each month. Earlier 30 GHz WVR and 9 GHz interferometer observations have demonstrated that winter nights give best phase stability; for this report we have therefore divided the data further between day and night to allow some comparisons to be made. The brief discussion below refers to average data.
Figure 1 shows the percentage of clear nights (loosely defined as 7 pm - 7 am) for the decade. At this point it is worth summarizing the results of the earlier site testing reports (e.g. AT/31.6.7/013). The often-reported 60-nights of 3 mm observing time occured during the period May to Sept. in 1992 (a fairly typical year in our present statistics). Note that there are a substantial number clear nights either side of this window.
Figure 2 shows the percentage of clear days. Note that over a quarter of all days in the "optimum" May to Sept. window are cloud-free.
Implications for the 225 GHz Clear-Sky Phase Correction Project
If we are fairly conservative and concentrate on a 6-month winter window starting in May, we may reasonably expect our WVR phase correction scheme to work well for 40% of the nights and 25% of the days: the equivalent of a total of 117 twelve-hour syntheses. Relative to the previous estimate of 60 nights of uncorrected 100 GHz synthesis, we are therefore very close to the working approximation that the WVRs will double the amount of 100 GHz observing time.
The overall situation is rather better than this however. We strongly suspect that given the clear periods available, phase correction may be used to extend usable 7 mm synthesis time out to perhaps nine months of the year. Finally, we note that the scheme will make worthwhile contributions to 1 and 3 cm synthesis around the summer season.
In formulating these estimates, we note that dynamic scheduling of the ATCA time is important if maximum advantage is to be gained from the phase correction system. This obviously becomes more important as one departs further from a winter night mm-wave synthesis schedule.