CSO Tau

The CSO Tau Monitor measures the 225 GHz opacity every 10 minutes, more frequently than the rate at which skydips are taken with the JCMT, albeit at a fixed azimuth. Comparing the skydip values to the CSO Tau Monitor values yields relations between Tau_CSO & Tau₈₅₀ and Tau_CSO & Tau₄₅₀. If the scatter about these relations is small, then the CSO Tau monitor can be used to measure the opacity more frequently than the skydips, with no additional overhead.

As mentioned above, these relations have been derived already, but there is some worry about using them, especially given the discovery that the previous values of T_HOT and T_COLD were incorrect. The CSO Tau values also tend to spike up and down quite a lot, which may or may not be representative of the sky itself. In addition, the relations have only been calculated for the narrowband filters, and may be different for the new wideband system.

Before recalculating the relations for both the narrowband and wideband filter systems, taking into account the revised values of T_HOT, T_COLD and $\eta _{\rm tel}$, we wish to describe a new method of handling the CSO Tau data, that reduces the scatter in the CSO Tau relations even further.

Figure 3: High resolution CSO Tau data as a function of time (expressed as a fraction of the UT date) for two different nights. The data is depicted by the `+' symbols. The 850-micron skydips taken for each night, scaled to CSO Tau values using the newly derived CSO Tau relations (Section 5), are denoted by `*' symbols. The solid line is a polynomial fit to the CSO Tau data.

Currently, ORAC-DR reads the CSO Tau from the data-file headers. However, the CSO Tau Monitor may update more frequently than the rate at which observations are taken. We extract the Tau directly from the CSO archive, and plot this `high resolution' CSO Tau data as a function of time. A polynomial is then fit to the data to track the large scale variations in CSO Tau as opposed to the small-scale noise.

The high resolution data for two different nights are shown in Figure 3. There are several points worth noting:

1.

Using the relations derived in Section 5, we have also plotted the CSO Tau predicted by the 850-µm skydips taken each night. It is striking how well the skydips track the CSO Tau, even when one would imagine that the CSO was moving around too much to be useful.

2.

The high resolution CSO data show a significant amount of noise, but do track long time-scale (~2 hour) variations very well. This noise is more than one would expect given the measurement errors quoted in the CSO Tau archive. It is possible that this small-scale noise is due to instrumental noise and does not represent the behaviour of the sky. This is supported, but not proven, by the good correlation between the skydips and the polynomial fit. From experience at the telescope, we have noticed that sometimes the CSO Tau fluctuates, while the skydips indicate a uniform sky.

This work seems to indicate that the polynomial fits (i.e. the `smoothed' CSO Tau data) may be the best representation of the 225 GHz optical depth. Work is currently in progress to archive the polynomial fits and make them available to the general user community.

A note of caution, however - although in general the polynomial fits are very good, on some nights (especially when there is a large level of scatter and variability in the CSO Tau data) they may not provide an accurate description of the 225 GHz optical depth. Thus, the smoothed CSO Tau data should not be used blindly nor taken as a replacement for performing skydips. Version 1.1 of ORAC-DR supports the polynomial fits if requested on the command line.