Tomoe Note 017: Analysis of 20191124 = Nov 24, 2019

Michael Richmond
May 29, 2020

Executive Summary

I have analyzed a dataset involving all 84 chips. The run is interrupted in the middle (by clouds?), and so contains only 32 minutes worth of data. No promising candidates were found, but three of the candidates were faint (known) asteroids, and one is a faint star with high proper motion.


The dataset for 20191124

The dataset "20191124" is based on images acquired during UT 2019 Nov 24. The images cover the span of time



    2458812.1066740  ≤   JD  ≤ 2458812.1365636

which is about 32 minutes.

The FITS headers of one of these images states, in part:

OBJECT  = 'J0354+1749_dith1'   / object name                                    
EXPTIME =           119.988480 / [s] total exposure time                        
TELAPSE =           120.500020 / [s] elapsed time                               
EXPTIME1=             0.999904 / [s] exposure time per frame                    
TFRAME  =             1.000000 / [s] frame interval in seconds                  
DATA-FPS=             1.000000 / [Hz] frame per second                          
DATA-TYP= 'OBJECT'             / data type (OBJECT,FLAT,DARK)                   
OBS-MOD = 'Imaging'            / observation mode                               
FILTER  = 'BLANK'              / filter name                                    
PROJECT = 'Earth Shadow Survey (1Hz)' / project name                            
OBSERVER= 'Noriaki Arima'      / observer name                                  
PIPELINE= 'wcs,stack,raw'      / reduction pipeline templete                    

The images were reduced and cleaned by others; I started with clean versions of the images. Each set of 120 images was packed into a single FITS file, covering a span of (120 * 1.0 sec) = 120 seconds. These "chunk" files were located on shinohara in the directory

   /gwkiso/tomoesn/raw/20191124 

with names like

   rTMQ2201911240018612114.fits

These names can be decoded as follows:


       r             stands for    "reduced" ??
       TMQ2          means         "Tomoe data, part of quadrant 2"
       20191124      means         year 2019, month 11, day 24
       00186121      means         chunk index 00186121 (increases with time)
       14            means         chip 14
       .fits         means         a FITS file 

I'll refer to each of these "composite" files as a "chunk".

There are typically 15 chunks for each chip, and a total of 1245 chunks in the entire dataset. Each chunk file was 1083 MByte, so the total volume of the chunk files was about 1348 GByte = 1.3 TByte.


Brief description of the pipeline analysis

I ran a slightly modified version of the Tomoe pipeline on the images; it was not the same as that used to analyze the 2016 images discussed in the transient paper for two reasons:

  1. minor modifications to handle the slightly different format of the data (image size and overscan regions)
  2. small improvements in the astrometric procedures to make the routines more robust

The main stages in the pipeline were:

  1. split a chunk into individual images (120 images per chunk)
  2. detect and measure stars in each image
  3. add the Julian Date to each stellar measurement
  4. perform ensemble photometry of all images in the chunk
  5. perform astrometric and photometric calibration of the stars included in the ensemble (some stars, detected in only a small number of images, will not be included in the ensemble)
  6. perform astrometric and photometric calibration of all stars detected in each image individually

The output of the pipeline includes a copy of each FITS image, plus a set of ASCII text files which include both the raw, uncalibrated star lists, and the calibrated versions of those lists, as well as the ensemble output.


"Weather" plots for the 4 quadrants

After running the pipeline to reduce the data, clean the images, find and measure stars, calibrate them astrometrically and photometrically, I used a script to look at properties of the data over the course of the night. You can read more about the "weather" in another note.

Below are links to the graphs produced for each of the 4 quadrants.

Quadrant 1:

Quadrant 2:

Quadrant 3:

Quadrant 4:


What do we learn from the "weather" graphs?


Transients and control time

After all the data had been calibrated, I ran the "transient_a.pl" script, which applies the rules described in the Tomo-e transient search paper to look for sources with only a brief existence. The code also computes a "control time" for the dataset.

The software found 45, 80, 37, and 32 candidates in quadrants 1, 2, 3, and 4, respectively. I created a web page showing the properties of these candidates in each quadrant:

The entry for each candidate includes some information about the chunk in which it appears, its position in (x,y) pixel coordinates and (RA, Dec) coordinates, and its magnitude. The "variability score" describes the ratio of the standard deviation of its magnitudes away from the mean to the standard deviation from the mean of stars of similar brightness; so, a high score means the object is varying from frame to frame more than most objects of similar brightness.

The entries for quadrants 2, 3, and 4, contain columns listing the magnitudes of any objects at this position (to within 5 arcsec) in the USNO B1.0 (avergage of R-band magnitudes) and in the 2MASS catalog (K-band magnitude). A value of "99.0" indicates that no source appears in the catalog as this position. You can see that the overwhelming majority of candidates do correspond to objects which were detected in one or both of these catalogs -- meaning that they are not true transients.

After these columns of text, the documents contain thumbnails of the images around the candidate. The thumbnails are oriented with North up, East left, and are 110 pixels (= 130 arcsec) on a side.

Interesting candidates are:

The table below shows the control times for each quadrant in this dataset:


quadrant         control time (square degrees * sec)
                   V=13     V=14     V=15     V=16    V=17  
-------------------------------------------------------------
  1                8457     8457     8457     7565    1380
  2                8457     8457     8457     7766    2301
  3                8054     8054     7996     6370     201
  4                8021     7995     7743     5803       0

 total            32989    32963    32653    27504    3882
-------------------------------------------------------------

These control times are comparable to those for each of the nights listed in the transient paper;