In a meeting of the TVS Collaboration today, several questions came up:
(1) PSTN-055 indicates that the WFD cadence would be one observation in one filter followed by a break of 33 minutes and then another observation in another filter. PSTN-056 indicates that two 15 second snaps will be combined into one 30 second snap. Do both of these statements still stand? That is, will there be one 30 second snap in one filter followed by a break of 33 minutes and then a second 30 second snap in another filter? Or, is there just one snap of 30 seconds in one filter for each pointing for each day?
(2) I have previously understood from this forum that the average on-sky time each day will be 4.7 hours. Is this on-sky time only for the WFD survey with additional time allocated for the mini-surveys including the DDF? Or, on average is 4.7 hours the total of all observations for all surveys each day? I understand that this is the average and that some days will be longer and some shorter.
Pinging survey strategy/cadence expert @ljones for any further input.
PSTN-056 says in Section 7: “In all of these regions on the sky, visits within a night are attempted in pairs, with each first visit to a pointing paired with a return visit typically 33 minutes later in a different filter” which I read as saying that in the baseline strategy of the latest SCOC recommendations, visits to a given pointing are still generally planned to be paired with 33 minute spacing. Spacing visits to a given pointing by 33 minutes can happen regardless of whether each visit is split into two 15 second snaps or done as one 30 second long exposure.
Aaron, I got the 4.7 hour number from Lynne Jones in a previous post (Jan 6 Background on the external conditions for Rubin - #6 by sdcorle1)
Yes, that’s still the general plan - two visits for each field, separated by 33 minutes, and those visits are still nominally 2x15s (30s on-sky) exposures.
It is nowhere near official, but it’s highly likely that instead of a visit being 2x15s, that it will instead be 1x30s.
This by itself will increase our on-sky efficiency, by somewhere close to 10% … that increases the average on-sky exposure time from 4.7 hours to a bit over 5 hours.
The average on-sky exposure time I mentioned in that other post really is the total amount of actual on-sky exposure time, for all purposes (I didn’t split out WFD vs. DDF).
However, I would just like to be clear that when we talk about visits being separated by 33 minutes, this is a different “kind” of time than the 4.7 hours. The 4.7 hours is made up of 30s exposures, followed by readout, slewtime, filter change time, etc. – so each of the visits is separated (on average) by about 7 seconds from the previous exposure. The total amount of time spent “doing observing” in each night is longer than 4.7 hours.
Thanks. I’m glad that you clarified that the 33 minute gap between visits is still in the picture. I, also, think that my current planning should include the 7 second delay.
Based on the “pure” observing time of 4.7 hours, I understand that there would be 564 of these per night. The 564 multiplied by 30 seconds give you the 4.7 hours. If I add 7 seconds to each pointing, that gives a total of about 5.8 hours of elapsed observing time. Does that seem like a reasonable average nightly time? (I understand that nights with good seeing will be longer and those with bad weather/full moon etc. may be wash-outs.)
It’s a little longer than that - the mean length of the night available from -12 to -12 twilight is about 10 hours, however the mean of the “actively used” hours is only 6.5 – this does include slewtime (which has a median time of ~7 seconds, but there are also some long tails and filter changes take considerably longer, at 120 seconds).
So it’s closer to 6.5 hours of observing time per night, after deducting weather or engineering time.
Looking at the opsim output database v4.3.1, I see an average of 727 visits per night, for nights with any observations (although this does translate to an average of 552 visits per night for all nights during the survey including those with no observations at all).
I suppose I should also note that DDF visits do not get a second visit at 33 minutes (they are instead observed straight through in a sequence that lasts something like an hour), neither do some types of visits in subregions of the sky … but the vast majority do have pairs at ~33 minutes.