search

thespacedoctor / sherlock

The QUB Transient Classifier
GNU General Public License v3.0
11 stars 5 forks source link

Sherlock and the SMC and LMC #149

Closed thespacedoctor closed 2 years ago

thespacedoctor commented 2 years ago

I'm investigating the current matching radii for the SMC and LMC as determined by Sherlock.

For both galaxies, Sherlock uses the semi-major axis dimensions (from NED-D) alongside a "galaxy radius stretch factor" to calculate a transient association radius. On close inspection, these radii look to be a little too generous (see below). I'll post suggested solution(s) here.

SMC

RA 00:52:44.8
DEC -72:49:43.0
Morphology: 320x185 arcmins

Association radius = (320 / 2) x 1.2 = 192 arcmin

LMC

RA 05:23:34.5
DEC −69:45:22.0
Morphology: 645 x 550 arcmins

Association radius = (645 / 2) x 1.2 = 387 arcmin

thespacedoctor commented 2 years ago

I'm starting to develop a generic solution for the angularly very large extragalactic hosts (instead of hard-wiring exceptions for LMC and SMC).

Here are all sources in NED-D with major-axis diameter > 50 arcmin with screenshots below:

primary_ned_id major_diameter_arcmin dist_mpc
2MASX J19303773-5217259 60 124
MESSIER 033 70.8 0.692
MESSIER 031 190 0.745
SMC 320 0.063
Sagittarius Dwarf Spheroidal 450 0.022
LMC 645 0.0535

M31

Sagittarius Dwarf Spheroidal:

2MASX J19303773-5217259

This is worrying as it has obviously incorrect dimension data in NED (60 arcmin major axis!):

M33

M31

Sagittarius Dwarf Spheroidal:

2MASX J19303773-5217259

This is worrying as it has obviously incorrect dimension data in NED (60 arcmin major axis!):

M33

Suggested Actions:

A longer-term enhancement might be to use full galaxy geometry to produce elliptic search radii based on major & minor and parallactic angle (don't have PA in database yet but can harvest from NED).

thespacedoctor commented 2 years ago

Changes are in and will be pushed in the next release.

ENHANCEMENT: added separate treatment of galaxies which extend > 10 arcmin in the sky (65 sources in total) Search radii are reduced to reduce the chance of a background source getting incorrectly associated with these local neighbours. ENHANCEMENT: added a 'deny list' in settings for galaxies with incorrect morphology reports in NED (only morphology ignored, all other metadata read as usual). So far list members only include "Sagittarius Dwarf Spheroidal" and "WISEA J193037.70-521726.0" (aka 2MASX J19303773-5217259)

Here you can see the before (larger) and after (smaller) search radii for our 4 largest neighbours.

LMC

SMC

M31

M33

thespacedoctor commented 2 years ago

Sherlock v2.1.1 has been released