Continue the Cycle 17 NUV and FUV dark programs (11894 and 11895) essentially unchanged. They obtain two NUV and five FUV darks each week.
There are two general uses for these data: First, we need to monitor the overall dark rate as a function of orbital position and of time since launch. I believe that Tom Ake wrote tools to do this. I assume that Wei has access to them. If we could automate this process, then we could generate a dark-rate map each month and put it on the COS monitors page. Another useful plot would be the mean count rate in some region far from the SAA as a function of time since launch. The RIA would not have to interpret either plot, except to check that the dark rate is following the expected trends.
Second, we will eventually want to construct reference dark frames that we can scale and subtract from the data. At this point in the mission, we probably have enough dark frames that we can begin experimenting with how best to do this. For example, Dave Sahnow suggested that we use the cumulative-image routine to construct a summed dark-frame image each month. Perhaps OPUS could do this automatically; perhaps an RIA should do it. (Perhaps the cumulative-image routine is not the best tool for the job.) It might be worth investigating whether the distribution of dark counts on the detector varies with orbital position. We certainly expect structure to appear in the dark frame as gain sag becomes important, so eventually want a set of time-dependent reference darks. Again, the RIA would make these files, while an instrument scientist would interpret them.
Once we figure out what the reference dark frames should look like, then we could ask an RIA to construct, analyze, and deliver them to CDBS on a regular basis.