Rubin2020_PCW slack archives day3-wed-slot2a-survey-strategy 2020-07-15---2020-08-17

Wed 2020-07-15 03:02PM

@Ranpal (she/her/hers) has joined the channel

Thu 2020-07-16 01:19PM

@Melissa Graham has joined the channel

Wed 2020-07-22 11:57AM

@Jeff Carlin has joined the channel

Wed 2020-07-22 01:17PM

@Lynne Jones has joined the channel

Thu 2020-07-23 05:35PM

@Peter Yoachim has joined the channel

Lynne Jones Thu 2020-07-23 05:39PM

Let's talk survey strategy!

Wed 2020-07-29 07:57PM

@Preethi Nair has joined the channel

Thu 2020-07-30 03:43AM

@Christian Setzer has joined the channel

Fri 2020-07-31 03:04AM

@Ilaria Musella has joined the channel

Fri 2020-07-31 08:32AM

@Robert Szabo has joined the channel

Mon 2020-08-03 11:40AM

@K-T Lim has joined the channel

Mon 2020-08-03 07:28PM

@Ian Sullivan (he/him) has joined the channel

Lynne Jones Mon 2020-08-03 09:11PM

We are trying to plan for discussion during this section: we will have videos available to pre-view, a quick summary of those videos (SUPER quick, it won't be a substitute but a reminder), and then short responses from the various science collaborations.

Lynne Jones Mon 2020-08-03 09:13PM

One resource will be the report we are preparing for the SCOC - https://pstn-051.lsst.io/ ( https://pstn-051.lsst.io/v/v1.0/index.html is the specific draft we have sent to the SCOC on 8/3/2020).

Lynne Jones Mon 2020-08-03 09:13PM

We will add links to our videos as we get them uploaded.

Lynne Jones Mon 2020-08-03 09:14PM

And, as always - http://community.lsst.org is a great place for discussion as well : https://community.lsst.org/c/sci/survey-strategy/

Wed 2020-08-05 07:45AM

@Tomás Müller has joined the channel

Wed 2020-08-05 10:45PM

@Luis Pineda has joined the channel

Lynne Jones Thu 2020-08-06 04:23PM

Summary of SCOC schedule and membership https://www.lsst.org/sites/default/files/SCOC%20Handout.pdf

Thu 2020-08-06 06:10PM

@Kara Ponder has joined the channel

Fri 2020-08-07 12:19AM

@mrawls has joined the channel

Fri 2020-08-07 10:15AM

@Karen Masters has joined the channel

Melissa Graham Fri 2020-08-07 02:15PM

@Melissa Graham set the channel topic: Evaluating Survey Strategies
<https://project.lsst.org/meetings/rubin2020/agenda/session/community-evaluation-rubin-survey-strategies>

Fri 2020-08-07 03:19PM

@Markus Rabus has joined the channel

Lynne Jones Fri 2020-08-07 04:51PM

Also watch this thread on http://community.lsst.org :
https://community.lsst.org/t/survey-strategy-session-at-the-rubin-obs-pcw-2020

Sat 2020-08-08 12:09AM

@Lee Kelvin has joined the channel

Sun 2020-08-09 07:14PM

@Michael Solontoi has joined the channel

Sun 2020-08-09 07:57PM

@rmandelb has joined the channel

Michele Bannister Tue 2020-08-11 12:15AM

Qn: are there any plans to run the survey strategies (e.g. twilight survey) with the scheduler playing dodge-a-megaconstellation?

Lynne Jones Tue 2020-08-11 03:30AM

We didn't refer to it in our talks, but we have tried a basic satellite avoidance simulation. The avoidance was as simple as "is there a satellite pass expected in the next visit? if so, delay the visit by a few seconds to allow the satellite to pass". The results are described briefly in the short talk by Tony Tyson (from the Monday sessions) - https://youtu.be/ulVRA3z7C1I
The upshot of this kind of simple avoidance is that we spent more time waiting for satellites to pass (over the lifetime of the survey) than we seem likely to lose in terms of pixel area, if they are not so saturated that they ruin large fractions of the image. Even if we lose a few percent of each image with a satellite, this still beats the time lost waiting.

Robert Lupton Wed 2020-08-12 01:47PM: I don't expect to lose large areas, although some of the early reports estimated rather scary areas. Only the basic trail is lost, which is a few arcsec across (it's out of focus), so a fraction of a percent.

Lynne Jones Tue 2020-08-11 03:33AM

That said, there are perhaps more complicated 'avoiding the satellites' games you could play. Part of that would depend on working with the companies putting up the satellites to get very accurate positions of their sats over the night (which they could do, just longer term predictions start to get a bit less accurate). We have some uncertainty in the pointings for the survey, because we don't know exactly when we're going to arrive at a particular field .. there can be slowdowns due to small mechanical issues, for example, that we don't know how to quantify yet. So let's say our queue for the next hour or two hours includes about five minutes of uncertainty (we will point at RA/Dec_1, at time t_1 +/- five minutes) .. with this information and an accurate orbit, we could figure out when we make that queue if we're likely to be looking at a satellite and reorganize the series of pointings so that we are more than 10 minutes away from crossing paths with one.

Lynne Jones Tue 2020-08-11 03:34AM

Well, maybe .. it depends how densely they cover the sky and how much that reorganizing would end up costing us in time (we try to cover the sky pretty efficiently, so adding a few seconds of slew time is actually a pretty big penalty if it happens often).

Lynne Jones Tue 2020-08-11 03:34AM

So whether or not this wins over just taking the area penalty is still TBD.

Lynne Jones Tue 2020-08-11 03:57AM

Please note that not all of our pre-recorded content has managed to find its way onto the proper web pages yet, plus I wasn't entirely sure how to get an updated/more detailed agenda there. However, the post on community contains everything!
https://community.lsst.org/t/survey-strategy-session-at-the-rubin-obs-pcw-2020

Melissa Graham Tue 2020-08-11 12:05PM: @Lynne Jones I see a lot of videos posted to your Survey Strategy page now. If you need the text updated I can do this for you. Let me know here or in #session-chairs
Lynne Jones Wed 2020-08-12 03:24PM: Thanks Melissa - do you think you can add a link to the community post in the agenda text?
https://community.lsst.org/t/survey-strategy-session-at-the-rubin-obs-pcw-2020/4305
Melissa Graham Wed 2020-08-12 03:44PM: @Lynne Jones added! https://project.lsst.org/meetings/rubin2020/agenda/session/community-evaluation-rubin-survey-strategies

Ranpal (she/her/hers) Tue 2020-08-11 06:04PM

Session:_Community Evaluation of Rubin Survey Strategies
Date/time: Wednesday August 12, 2020 - 07:30 HST - 10:30 PT - 13:30 EDT - 19:30 CEST - 03:30 AET +1
Connection:_ https://stanford.zoom.us/j/92715550713?pwd=S2Q0YTI4TmtHckVWRVBQT1JFZm5oUT09
Passcode: 257500
Details:_ https://project.lsst.org/meetings/rubin2020/agenda/session/community-preparations-early-science-rubin

Michael Strauss (he/his) Wed 2020-08-12 01:53PM

A general comment: a metric may say that a given survey strategy is 10% "better" than another. But the conversion between any metric and science output may be quite non-linear, with, e.g., thresholds below which science goals cannot be met, or very weak scaling of quality of science with the metric. I guess it will be the job of the SCOC to do those non-linear transformations as they decide on the survey strategy.

Will Clarkson Wed 2020-08-12 01:54PM: Agreed - setting the thresholds for "good / acceptable / unacceptable" is a challenge we're trying to overcome... Does the SCOC formally require thresholds?
Will Clarkson Wed 2020-08-12 01:55PM: ^^or perhaps asked a bit more clearly: are thresholds the best mechanism for the SC's to make it easier for the SCOC to recommend our science?
Michael Strauss (he/his) Wed 2020-08-12 01:56PM: I see that Meg, with her "green/yellow/red" approach is one way to think about this. Having said that, not all science is well-defined in terms of thresholds, and that's OK.
Meg Schwamb Wed 2020-08-12 02:02PM: I would argue that if someone wants to argue for a 10% increase they have to say why they can't do with it. I think that's the same whether you have a threshold or use the stop light analogy
Robert Lupton Wed 2020-08-12 02:06PM: Another problem with saying, "This is red" is that it makes it hard to tradeoff with other science, as no-one gets a veto
Meg Schwamb Wed 2020-08-12 02:07PM: Red means we don't recommend as a science collaboration because it doesn't satisfy our science, the SCOC gets to take that information and make that tradeoff
Robert Lupton Wed 2020-08-12 02:08PM: But it doesn't help with things that are actually orange! Like "ifneedsbe" in Doodle. Maybe it won't be a real problem
parejkoj Wed 2020-08-12 02:10PM: I would assume that's what "yellow" is for?
Meg Schwamb Wed 2020-08-12 02:10PM: yes
Meg Schwamb Wed 2020-08-12 02:11PM: And then when that gets identified by the SCOC, each SC and the community can give more feedback. I think it's very good that this will be an iterative/interactive process
Robert Lupton Wed 2020-08-12 02:13PM: I'd still argue for a number if [0,1] --- binning loses information
Colin Slater Wed 2020-08-12 02:16PM: cross calibration is a bigger challenge than information loss I think
Meg Schwamb Wed 2020-08-12 02:16PM: I agree. Thresholds some number will go into the stop light colors, but I think also recommending which simulations we think do best for Solar System science, where the families of simulations do fine and meet the criteria, and where they do bad will be helpful for the SCOC to digest the SSSC viewpoint.
Lynne Jones Wed 2020-08-12 02:34PM: I also think this approach is going to be generally likely to match metric results across the science collaborations - some kinds of runs are great, some are bad and most are 'not a lot different'.

Stephen Gwyn Wed 2020-08-12 02:01PM

On Meg's last plot, am I seeing correctly that most strategies are within 5% of each other for SS science? How does this relate to Michael's comment above?

Meg Schwamb Wed 2020-08-12 02:02PM: For that specific suite of families of simulations.
Meg Schwamb Wed 2020-08-12 02:04PM: It means most of the footprints as long as they have the northern ecliptic spur don't too bad for Solar System
Stephen Gwyn Wed 2020-08-12 02:05PM: Make sense, thanks.

Francisco Forster Wed 2020-08-12 02:08PM

A general question: are the new satellite constellations taken into consideration via any metric? If not, are there any plans to consider them? For example, one may want shorter exposures.

Robert Lupton Wed 2020-08-12 02:12PM

I think that the area lost to a full-focal-plane trail seems to be pretty small (0.2%??), and the cost to shorter exposures is large (min. 3s per readout, and maybe more if we decide to slow the read). That's not a reason not add a metric, but it'd probably also require more work on the scheduler to be interesting. Lynne mentioned that they'd tried delaying, but not e.g. switching the order of observing patches of the sky

Francisco Forster Wed 2020-08-12 02:16PM: I wonder if there are other change besides exposure that could be relevant, e.g., where you observe close to sunset/sunrise.
Peter Yoachim Wed 2020-08-12 02:17PM: yes, staying away from the horizon could significantly decrease the number of streaks. The issue is that is where folks would like to look for NEOs
Francisco Forster Wed 2020-08-12 02:25PM: Thanks!
Michele Bannister Thu 2020-08-13 04:44AM: NEOs, comets, interstellar objects...

Alison Rose Wed 2020-08-12 02:27PM

Was the metric hackathon recorded

Rstreet Wed 2020-08-12 02:27PM

The plenary sessions were yes

Rstreet Wed 2020-08-12 02:28PM

You can find the recordings linked from the hackathon website

Alison Rose Wed 2020-08-12 02:28PM

Thanks! Hi Rachel!

Rstreet Wed 2020-08-12 02:28PM

https://lsst-tvssc.github.io/metricshackathon2020/

Rstreet Wed 2020-08-12 02:28PM

Hi there @Alison Rose !

Alison Rose Wed 2020-08-12 02:29PM

Slack question: when should I be using my @ sign?

Rstreet Wed 2020-08-12 02:29PM: If you use the @ sign in front of someone else's name, they get a notification or a dot to draw their attention
Robert Lupton Wed 2020-08-12 02:30PM: If you type, "Robert Lupton is Wonderful!!!" I'll never know without an @
Peter Yoachim Wed 2020-08-12 02:31PM: and in Slack you can also put the @ in front of channel , here , or all I think, so e.g., everyone in the channel gets pinged
Robert Lupton Wed 2020-08-12 02:31PM: Please don't do that, though.
Robert Nikutta Wed 2020-08-12 02:32PM: And, depending on your everyone's setting, @-ing someone or a channel will send email(s) to them too.
Alison Rose Wed 2020-08-12 02:34PM: @Robert Lupton @Rstreet @Peter Yoachim @Robert Nikutta Thanks for the info! see you after the break.

Ferguson Wed 2020-08-12 02:29PM

BIG THANKS for the great summaries of all of the experiments and rationale behind them SCOC report...and for organizing this session!

Eric Neilsen Wed 2020-08-12 02:30PM

The lsst_sims software distributed in CVMFS is also very helpful, at least for those on Macs or Linux.

Lynne Jones Wed 2020-08-12 04:05PM: I have found the mac binaries to be useful too (at least when the build is working, which it occasionally does not).

Michelle Lochner (she/her) Wed 2020-08-12 02:30PM

thank you all!

Dave Young (QUB, UK) Wed 2020-08-12 02:30PM

thanks all!

Luis Pineda Wed 2020-08-12 02:30PM

thanks all

Robert Nikutta Wed 2020-08-12 02:30PM

Thank you!

Simon Krughoff Wed 2020-08-12 02:31PM

Whoa: https://platform.leolabs.space/visualization

mrawls Wed 2020-08-12 02:34PM: Nice, I hadn't seen this one
Simon Krughoff Wed 2020-08-12 02:35PM: Try ticking the "debris" box.

Alison Rose Wed 2020-08-12 02:32PM

@Mario Juric thanks for the leolabs.space link

Robert Lupton Wed 2020-08-12 02:32PM: An excellent use of @
Mario Juric Wed 2020-08-12 02:33PM: They provide real-time satellite tracking services (operate their own radars, etc).
Alison Rose Wed 2020-08-12 02:36PM: @Mario Juric @Robert Lupton I'm gearing up my work on LEO. I visited Tony before the lockdown, and kept in touch.
Mario Juric Wed 2020-08-12 02:37PM: I suspect they may be willing to let the astronomical community use their services for prediction/avoidance for a small fee (or possibly free). Ed Lu is one of their co-founders (and very active in supporting planetary defense research with the B612 Foundation).
Michele Bannister Thu 2020-08-13 04:51AM: Hmm, they have a station here in NZ now
Robert Lupton Thu 2020-08-13 09:38AM: Do I hear, "Thinks: In Kind??"?

Meg Schwamb Wed 2020-08-12 02:32PM

More info on the SCOC and timeline for decision making including upcoming community workshops

https://www.lsst.org/sites/default/files/SCOC%20Handout.pdf

Stephen Smartt Wed 2020-08-12 02:44PM

@Lynne Jones @Peter Yoachim left a message for you on community on the cadence report, thanks,

Lynne Jones Wed 2020-08-12 02:45PM: Can I move it to the cadence report topic - it's in the topic regarding the workshop session? (working on an answer)
Stephen Smartt Wed 2020-08-12 03:00PM: Sure, better there

Ranpal (she/her/hers) Thu 2020-08-13 05:32AM

The recording of the live session is here: https://youtu.be/wKzPcT4WmOo

Lynne Jones Thu 2020-08-13 02:24PM: Thanks @Ranpal (she/her/hers) ! FYI, I think the wrong video got linked into the session agenda page ( https://project.lsst.org/meetings/rubin2020/agenda/session/community-evaluation-rubin-survey-strategies )
Ranpal (she/her/hers) Thu 2020-08-13 02:32PM: thanks will get it fixed

Jovan Aleksic Sun 2020-08-16 06:20AM

@Peter Yoachim , @Lynne Jones , the slides od the talk " Survey Strategy: The 1.6 Set" are placed at the " Simulation Experiments" video. Move them at the corresponding one.
Also, are the slides of the remaining two talks available (Simulation Experiments and SCOC Survey Strategy)?

Lynne Jones Mon 2020-08-17 02:39PM

We can't edit the page, but I'll let @Emily (she/her) know - so the slides aren't aligned? There are slides for the 'Simulation Experiments' video, Emily should have those also so can probably sort out both of those at the same time. @Emily (she/her) could you please modify the agenda page so that the slide sets match with the videos?

Lynne Jones Mon 2020-08-17 02:39PM

I can upload the slides for the videos I made to community too.

Emily (she/her) Mon 2020-08-17 03:00PM

Fixed. But I don't see the SCOC Survey Strategy slides. Can you resend

Lynne Jones Mon 2020-08-17 03:44PM

Ah, sorry - the interview with Zeljko did not have slides. (he wrote the slides he talks about, but I don't have a copy of them).
Could you maybe just include a link to the SCOC webpage on http://lsst.org at the top of that talk? ( https://www.lsst.org/content/charge-survey-cadence-optimization-committee-scoc )