Little is known about the distribution of point sources at these frequencies, and without a dedicated program of followup with a high-gain (K/Jy) instrument, the best that can be done is to avoid known sources from low-frequency surveys. From the DASI data alone, we can identify and remove point sources to approximately 0.25 Jy, and generous assistance by the ATCA facility has provided us with additional observations of contaminating sources. The contribution from sources below the DASI detection threshold will require a statistical correction to the measured power spectrum at our highest-l bins, but we expect this to be quite small.
Astronomical foregrounds aside, the presence of ground contamination at a level much greater than the expected cosmological signal places by far the most stringent constraint on our observing strategy. DASI was intended to operate with a ground shield; this could not however be installed until the 2000-2001 austral summer, and was not in place while the data presented here were collected. The signature of the ground is ubiquitous even in the raw visibility data, and is strongly dependent on baseline length, falling sharply with increasing u-vradius.
Accordingly, instead of tracking single fields over the full azimuth
range, we observe sets of eight fields, each field in the set spaced
by one hour of right ascension. Fields are successively tracked over
two narrow azimuth ranges, selected to avoid lines of sight which
intersect with station buildings. In this way, during a 24-hour
period, we obtain two hours of data on each of eight fields, with the
remainder of the time spent on calibration. The number of fields
reflects a compromise between our ability to constrain the ground,
observing time on each field, and sample variance on the CMBR signal.