In the context of a possible SSSC twilight survey white paper, what are the constraints on non-sidereal tracking? TLS-REQ-0160 says only “The telescope shall follow non-sidereal objects […] to within Tel_Track_Error.” How rapid non-sidereal targeting is reasonable to expect?
– Rob Seaman, Catalina Sky Survey
That’s an interesting question, I haven’t heard much discussion of using LSST in non-sidereal tracking mode.
Is the sky background too high that you can’t look for streaks on regular exposures? Because then the exposure can still be used for things like the WFD SRD requirement.
Trailing losses depend on PSF, pixel scale, rate of motion, etc. When targeting solar system objects the trade-offs for selecting exposure time, non-sidereal rates, or “track & stack” mode parameters will vary from target to target. SSSC ToO science cases (e.g., in support of NEOCam, the northern NEO surveys, or urgent response to impactors or interstellar objects) might well prioritize detection of a faint object over other requirements, especially at acute solar phase angles, for instance. Such may benefit from non-sidereal rates, thus a question of what LSST is capable of achieving here. By contrast, asteroid survey mode observations will generally be sidereal (whether or not MOPS is involved). Streak detection, per se, is not the essence of the exercise.
I will check with Telescope & Site to see what the actual non-sidereal tracking capabilities are. At the best case, the dome moves at a max speed of 1.5 deg/s in altitude and 1.75 deg/s in azimuth, and the telescope mount can move faster. However, I doubt that would be recommended during an exposure. 
I wonder if LSST would be a non-optimal followup tool for many of these cases, given its very large FOV. I would have thought that unless an object is coming out of the sunward side or quite small (well… smaller than the set of objects currently found as immediate close approachers), LSST is likely to have already detected it when it was further away and moving more slowly (so there is a smaller uncertainty in the ephemeris prediction, as well as more time to do recovery, which might mean doing followup on other telescopes). It’s an interesting question and I am hoping that either I or some other members of the solar system science collaboration can have a look at the question of “how many objects will LSST detect that need immediate followup?”.
Non-sidereal tracking would have implications on the reduction pipeline too.
Thanks for checking the non-sidereal capabilities. For comparison the final community follow-up of 2008 TC3 was 450 degrees/day. A more typical delta rate would be tens of degrees per day. The TCS and camera subsystems would need to be able to work at non-sidereal rates, too.
The LSST FOV is large, but not “very large” by NEO standards. The CSS Schmidt is twice as big and ATLAS has a four times bigger field than LSST.
The specific context here is follow-up in support of / as part of a near-Sun twilight NEO survey, and also of NEOCam objects, themselves originating as near-Sun targets. Amy Mainzer emphasized NEOCam follow-up capabilities at the recent SBAG meeting and LSST might be wise to consider returning the favor. As you say, other telescopes will naturally follow-up LSST targets and developing an LSST follow-up capability would not replace the community resources.
A more general context is that LSST’s large FOV and ability to locate faint objects would be an excellent tool for recovering objects at risk of becoming lost for whatever reason. Northern targets move south, near-Sun targets move toward opposition. I don’t know how many LSST 3-day linkages will need immediate subsequent follow-up, but the solar system catalog is a community resource and LSST might benefit from providing assists, not just dunking the ball all by yourself
My reading of the call for white papers is that alternate survey cadences / strategies are welcome? An LSST twilight NEO survey would benefit from experience gained by the current NEO surveys. That could certainly include an alternate branch of the pipeline. Non-sidereal follow-up would likely represent a small fraction of these data, but the benefit to the NEO catalog would be disproportionately large.
Yes, alternate strategies are welcome.
Alternate pipeline processing needs to be described in the white paper.
I checked with T&S about non-sidereal tracking limitations:
The upshot is that the dome speed is the limiting factor, although there are acceleration effects on the mirror that need to settle before an exposure could start. I think it means that you could start non-sidereally tracking an object, wait for the acceleration effects to settle (while tracking), and then start an exposure (at least, nobody told me that was incorrect).