Monday, September 14, 2015

Classical vs. queue observing - a radio-turned-optical perspective

As a radio-turned-optical astronomer (I fully embrace the dark side now), I wanted to write about observing schemes. In astronomy, there are two main schemes to observing, classical and queue observing.
Edwin Hubble classically showing us how it's done
Classical observing
An astronomer submits a proposal for a certain number of nights at a telescope. Upon proposal acceptance, the astronomer is scheduled to run said telescope for a certain number of nights. On those nights, the astronomer has complete control of what the telescope looks at. This is the much older-school way to observe (think: Edwin Hubble).
Pros: spontaneity, ability to change observing strategies “on-the-fly”, a deeper relationship with the instrument.
Cons: one must travel to the site, which can be very remote, no guarantee that you will get any data (it could be too windy, too cloudy, an instrument could have had problems…) without the ability to make up for the lost time. And you are going to be completely flipping your sleep schedule.

Queue observing
The CARMA array at Cedar Flat, purely queue based
An astronomer submits a proposal to observe a certain set of sources. Upon proposal acceptance, the astronomer sets up the observing strategies / scripts and submits them to the telescope. Staff experts check the script, and submit the observations to a common queue to be observed in order of priority, optimizing the telescope itself to take the most data.
Pros: If you have a high priority project, you are very likely to get your data, without a dependence on the weather. Experts set up the telescope for you to optimize the telescope for your particular science goals, you do not have to travel to the remote site (in the case of Hawaii, this might also be considered a con…)
Cons: Basically the same as the pros of classical observing. You are not able to change your mind about sources, and you won’t have a personal relationship with the telescope when you are reducing your data.

For decades, the radio astronomy community has favored a queue observing format, partially because understanding an interferometer like the Very Large Array is highly complex, and usually takes many years to truly understand how the data go from acquisition to publishable. Because of this, only a few people were able to dedicate the time it took to have a personal relationship with the telescope. I served enough shifts at the CARMA array (RIP…) to understand the special pitfalls and quirks, but it took me years. In these cases, the queue observing is likely the best method. This is how ALMA works - they hire experts that have a personal relationship with the ALMA telescope to write your observing scripts and reduce your data for you. At the same time, something special seems to be missing.

The 6.5m Baade Telescope at Las Campanas Observatory
Magellan on the other hand is a pure classical instrument. If you are awarded time, you will be flying to Chile and staying at Las Campanas. Currently, I am spending time at a summer school for the IRAM 30m, a single-dish millimeter telescope located in the Sierra Nevada in Spain. The 30m is a hybrid, where you propose for sources which go into a queue, but also has an on-site option, which is currently how I am using it. In both my recent Magellan run as well as this current IRAM run, I am really glad we were classically observing, because the spontaneity was essential. At Magellan, I looked at my final source of the run, and realized that it was not as interesting as I thought it was, so I dropped it. So all of a sudden, I had 2 free hours of 6.5m time and needed something to do. I took to Twitter and Skype to chat with my collaborators, and we decided that it made a lot of sense to get a spectrum of a mysterious clump located to the northwest of the nucleus of the galaxy I had already observed, so we to get a spectrum of the clump, requiring a longer integration. Because we had that flexibility, it is possible I found something new and novel.

At the 30m, a similar thing happened. We are observing CO lines to try to determine the redshift of a submillimeter galaxy. We had a few guesses as to where the lines might be, but were blown away at how bright it was when we found the correct redshift. Instead of taking an hour to get the line, we needed only 15 minutes! So, what to do with the last 3 hours of the night… if this was queue observing, we would have had one amazingly high signal-to-noise spectrum of one line. Instead, we zipped back to our other possible lines, switched the observing strategy to detect more lines in the forest, and even attempted to detect other fainter lines (and succeeding). Instead of one “the redshift of this thing is z=XX”, we now have the possibility to publish a paper with a CO SLED, and astrochemical analyses from the additionally detected lines.
Me in front of the IRAM 30-meter single dish in the Sierra Nevadas in Spain
In an era where we are seeing classical observing facilities go extinct in favor of queue observing, I think we need to recognize what we might be losing. As a radio astronomer who was raised as a queue observer, it was my experience on a radio telescope classically observing that has motivated my call to keep the option of classical observing - you never quite know what you are going to get!

No comments:

Post a Comment