Notes from a discussion with Stella Kafka of AAVSO

I just had a discussion with Stella Kafka from the AAVSO about some pretty exciting connections LSST can make with the community of non-professional astronomers via the LSST alert stream.

The main use case, and one I don’t think we support, is for transients that saturate in LSST. They will be right in the sweet spot for many amateur astronomers to follow up while saturated in LSST. A very cool opportunity from the EPO stand point is that when a transient saturates, we can put it in a special place to let people know it is available for follow up and even indicate when people begin taking data on that object.

Am I right that there is no way to trigger on an event where the object has reached saturation? Since we don’t publish the SFM data with the alert, there are no saturation flags. You could presumably look at the postage stamp, but that seems more difficult than it should be.


We have 64-bit flag as part of DIASource table, and thus part of the Alert.
Isn’t that good enough?

Btw, what’s “SFM data”?

SFM == Single frame measurement

My worry is twofold: 1: the flag won’t contain saturation information (I could be wrong about that if the bits are copied in from the measurement record), 2: We will throw out the DIASource as spurious if it saturates.

Never heard of SFM before (I know about “Single visit processing”,
and Source measurements…).

DPDD doesn’t say anything about flags, but saturation ought to be in there!

All DIASources will be included in the Alert stream. Why do you say it will
be “thrown out if it saturates”? Where does that come from?

I don’t know that it will be thrown out. We will get a lot of false detections around saturated objects because the PSF model will not be correct for them. We know that and will typically be able to mask those false detections. Maybe you have a smarter solution in mind, but that’s the simple thing to do. I’m just saying that if a true transient saturates, we will have to be careful not to flag it as a false detection.

Thanks for passing along the idea! I’ll add it to our discussion points once Tim Axelrod is back in May.

Please could you clarify one thing for any other amature astronomers who may find this thread: At what magnitude are objects expected to saturate LSST - so amatures may only get a ‘bright object’ alert rather than any light curve ?

I had a look at the science data which quotes a sensitivity of around mag 25 and a dynamic range of 18 bits = 13.55 magnitudes. My estimate is around mag 12-13 from this but it might be even less.

One thing to consider if you are intending to have saturation alerts is that there might be pre or post saturation measurements, how is the transition from saturated to within dynamic range going to be handled ?

The expected saturation points at the median 0.7" seeing (according to the exposure time calculator) are u, g, r, i, z, y = 14.7, 15.7, 15.8, 15.8, 15.3, 13.9. This is from Section 3.2 of the LSST Science Book; that estimate is a bit old, now, but I don’t know of any official update to it and I have no reason to believe it’s changed significantly.

In the yearly data release processing of the static sky, we should be able to make good measurements on sources that are only slightly saturated, and we’ll try to measure essentially everything, while setting warning flags indicating various levels of trustworthiness on objects with saturated pixels.

The question of how to handle them in the nightly difference image processing is much trickier, and I’ll have to defer to someone else in the project - though I think it’s quite likely we don’t have firm plans there yet, as it could be contingent on how well some algorithms that are still being developed work.

Thanks for your response - lots of very interesting and accesible information in there.

One thing I could not see is what is the stratergy for dealing with artificial satellites and space junk ? With the sensitivity of the system you will capture a lot of the orbiting junk both in Low Earth Orbit and in Geo Stationary Orbit.

Even my ‘small scale synoptic survey telescope’ (a Canon 10D Camera with a F2.8 18mm lens) managed to capture a reflection from a defunct satellite with a negative magnitude last year. They will be a big challenge to your object following algorithms as they move quite fast and can vary in brightness drastically as in my example above and of course Iridium flares though they tend to be extended objects rather than points.

Will the software detect meteors or will these get rejected as not being point like ? Though of course you get the occasional Meteor heading straight for the telescope that will appear point like.

Looks like you have a big challenge to identify and reject these sort of events…

I just noticed that the November meeting of the AAVSO has as its theme “AAVSO’s Role in the Age of Large Surveys.” Has anyone from LSST been invited to speak, or planning to attend?

Just a note to say that people who find and read this thread my also be interested in Identifying Suddenly Saturated Objects in the Prompt Pipeline