Discoveries using the telescopes on Mauna Kea

[TomK]

Paul,

An interstellar comet isn't going to necessarily be moving faster than one that formed in the solar system although it's likely it will. The reason we think this is an interstellar visitor is due to the shape and inclination of its orbit and it's moving faster than you would expect for a comet that formed in the solar system. A combination of the shape and angle it's coming from plus its speed suggests it couldn't have formed within the solar system. It essentially came from somewhere else and once it reaches its closest point to the sun (about twice the distance of the earth from the sun), it'll head-off again into interstellar space.

PS. The speed of a comet like Borisov is pretty much a moot point when it comes to damage it would cause if it hit the earth (it won't, by the way). The nucleus of Borisove is believed to be at least a mile in diameter. Something that large hitting the earth would be devastating.

[PaulW]

Thanks for the explanation Tom.

https://www.bbc.com/news/science-environment-49962134

Saturn pulls ahead in the moon race.

[TomK]

Thanks, Paul. I was going to post about the Saturn moons story but you beat me to it! The work was done using the Subaru Observatory on Mauna Kea.

In the meantime, a new planet-hunting instrument has been commissioned at Gemini that will search for potentially habitable exoplanets:

"Nearly a decade in the making, exoplanet-hunting instrument installed in Hawaii "

https://phys.org/news/2019-10-decade-exo...awaii.html

"Imagine trying to assemble one of those huge LEGO sets, except there's no instruction book; you've done it once before, but then you had to take it all apart and put it in little bags," said Jacob Bean, associate professor of astronomy and astrophysics at the University of Chicago. "Also you're at 14,000 feet, and when the air is that thin it impairs your judgment and thinking, and so here you are working 12-hour shifts lifting heavy things but also trying to put together a delicate instrument."

I've been involved in commissioning instruments delivered from overseas before on Mauna Kea, and know very much how Bean feels!

[randomq]

Tom, are you familiar with any of the programming/systems work that goes on behind the scenes? What programming languages, frameworks, big data backends, OS, etc are popular or up-and-coming for astronomy on Mauna Kea?

Python/R/Scala/Go? Hadoop/Hive/Spark? Debian/Centos/XP? Thanks.

[dan d]

U can allways join us at our WHACOS meeting first tues of the month. At keck waimea 7 pm.

Lotsa smart astro peps in attendance. Feel free to join us on Yahoo whaco forum
Aloha


HPP

[TomK]

Randomq - I can't speak for all the observatories, I'm not intimately involved in all of their systems, but do know quite a bit about UKIRT and JCMT's systems plus I have some knowledge of what the others use.

For general science work, instrument and telescope control plus data reduction and analysis, it's all Linux-based. I believe some observatories use Debian but we use CentOS. We may, in the future, switch to Red Hat in the future as it comes with support but obviously costs money.

EPICS (https://en.wikipedia.org/wiki/EPICS) is generally used for instrument and telescope control along with various in-house software. As for data reduction and analysis, Python is very popular these days. Much of our own software is now being written in Python but we still use an awful lot of older code much of it written using Perl, C, and C++. There are even some older programs written in Fortran which still do the job - indeed I still rely on some Fortran code I wrote back in the 90s for data analysis - it's still a good language for math-based tasks.

For database stuff, SQL and MySQL tend to be used.

Hope this is helpful.

[randomq]

Thanks, very interesting, had not heard of EPICS before. I think the last time I worked on Fortran I was submitting it in batch to the university mainframe in the 90s! Love Perl though; too bad it is falling out of favor. Almost surprised about MySQL, given all the big data hype and the presumption that you guys must collect a *lot* of data.

What kind of analysis would you do with your code? Looking for presence of certain elemental spectra? Shapes? Temporal patterns? Counting? Is it a lot of image analysis? Super-resolution / Interferometry?

Cool to see how programming makes these discoveries possible!

[TomK]

randomq - the observatories tend not to do much of the analysis, the expectation is for the team that applied for telescope time will reduce and analyze their own data. The database stuff is more for associating the data with those groups and providing all the other information they need, e.g., weather conditions, image quality, the calibrations they need etc.

If you're interested, you can see our system:

http://omp.eao.hawaii.edu/

but of course, most of it is private. The OMP is a mixture of MySQL and Perl.

For the wide-field data we take, IOW, survey data, then we actually send the raw data off to Cambridge University in the UK who will process the data and then they send the reduced data to Edinburgh where the data are made available to the teams involved in the survey work and after a proprietary period (generally a year), to the public.

We do a preliminary reduction of the data as they are taken and this lets us know in real-time if the data are acceptable (this also goes into the database) so we know that the data will likely pass quality checks, but to process these data properly takes a long time and is not possible to do in real-time, so that's where the other groups come in.

More stuff here if you are really curious:

http://www.ukirt.hawaii.edu/observing/co...kbook.html

When we are in so-called Cassegrain mode, i.e., not wide-field mode, then we expect the astronomers to come out and carry out the observations or at least observe remotely. In those cases, the teams can get both the raw data and usually fully-reduced and publishable data straight from the telescope as we are dealing with much fewer data and our in-house software does a really good job of providing fully-reduced data.

One question I think you're asking is if the software can provide real-time information on what we're detecting. In most cases, no, it really needs a scientist to sit down and look at the data. On the other hand, some things can be automated, especially in wide-field imaging mode. for instance, moving objects such as asteroids and comets can be detected using the software, or basically, any significant changes in an object's brightness.

Essentially, automation can detect changes, but it can't yet tell you if we're detecting water or other chemicals and atoms; that needs the expertise of an astronomer.

[TomK]

randomq - thought you might find this interesting - "Most Popular Programming Languages 1965 - 2019"

https://youtu.be/Og847HVwRSI

[randomq]

Thanks, I had no idea Ada and Pascal had been so popular. Now that Python is at the top, it seems like it must inevitably fall to one of the new challengers...

We should start teaching programming and astronomy in grade school, and get kids interested in STEM while they are still eager to learn! Of course we have to have some STEM jobs waiting for them here so they have something to look forward to aside from leaving for the mainland. All my younger coworkers talk about is how they worked on the mainland after high school, or how they plan to go back for a better job, once they save enough for moving expenses etc. Depressing.