Sunday, February 21, 2016

Random Topics: GW GRBs, mini-JWST, Exoplanets and lunch.

So after all the excitement last week, the subsequent week has been filled with discussions and other work getting caught up. For me, the last few days have been spent getting caught up on some review work that is going to be due soon, so this will be a fairly short post.


On Friday, for our high-energy astrophysics meeting at UMD, we had a nice round-table discussion with one of the Fermi team regarding the Fermi-signal coincident with the LIGO detection GW150914. Their arXiv (not-yet peer-reviewed) paper is here, and discusses their detection.

Fermi-GBM detected a signal with a fluence of $10^{49}$ ergs/s  between 1 keV and 10 MeV, lasting about a second. The fluence, duration, and hardness was similar to a weak short-GRB. The reason it was so poorly localized was that the Fermi spacecraft was pointed in orthogonal to the optimal direction to localize the signal source. They characterize this as a roughly 2-3 sigma detection, with a 0.0022 false alarm probability, though there are some details of this FAP calculation that I don't quite follow (particularly, why this scales with time from the GW trigger event).

It is worth noting that Fermi expects to see events of this significance randomly every few thousand seconds or so, therefore there is a possibility this is just a coincidence. SWIFT-BAT did not detect anything, though it was also pointed in the wrong direction and doesn't have as wide field of a detector, so they had to make do with observations 2 days after the fact, during which there was no detection. More interestingly, the INTEGRAL-ACS which has a much larger collecting area than Fermi-GBM, (it's a coincidence shield that surrounds the spacecraft) but also much higher background, did NOT see anything at the same time as the Fermi-GBM detection. However, they are sensitive to a different energy band, and if the short-GRB like spectrum is taken as a given, INTEGRAL is expected to miss about 50% of the bursts Fermi catches at that fluence.

This of course did not stop many many so-called "creative" ideas from flooding the arXiv about how a BH-BH merger produced an electromagnetic GRB-like counterpart. I won't get into them here, and will leave the discussion for another day.


This isn't really all that research related but I just wanted to post a picture of our new JWST model we just got from MESAtech. The James Webb Space Telescope is of course the very expensive IR space telescope that carries the hopes, dreams, and opportunity cost of the entire astronomical community into the future. It will be revolutionary for extragalactic astronomy, mapping of the local group, and exoplanetary atmosphere studies, among other things.

This 1/40th scale model was 3D printed as a reward for kickstarting MESAtech's fully robotic model for high-school outreach. The pieces were all snap together and Vicky assembled it in about 45 minutes, with just some wire snips and a few pieces of tape to help tighten up some fits. We got it for $50 during the kickstarter, but they are now available for purchase from the MESAtech site for $75. It folds/unfolds, and at 1/40th scale is just about the right size compared to LEGO minifigs.

Exoplanets and Lunch

Lastly, this week I organized the inaugural Exoplanet lunch at UMD, in order to bring together Drake's group with some other exoplanet researchers, like myself. Personally, I'm hoping that a closer interaction with the exoplanet observers will lead to some interesting projects that will help me get a better feel for the details of exoplanetary data analysis and observations. Towards that end, I'm considering running a brief tutorial over the course of several lunches exploring the use of Dan Foreman-Mackey's emcee code for doing parameter estimation using Markov-Chain Monte-Carlo methods. Jake van der Plas has a very nice tutorial he ran at the last AAS, which would be a very good basis to start from. 


  1. I just started playing with emcee last week. It's pretty snazzy. You should summarize your tutorial into some remarXiv posts.

  2. Check out Jake's tutorial on github. It's really good.