GRB221009A src-ind analysis dark conditions

Back to the Data analysis page

Go back to Transient Working Group.

Go Back to Gamma-Ray Bursts (GRBs).

Go back to GRB221009A.

• Analysis by A. Aguasca-Cabot (Universitat de Barcelona - arnau.aguasca@fqa.ub.edu)

General information

This wiki web page shows the analysis of GRB221009A data recorded in dark conditions (run_id: 9625-9960) and in mild moon conditions (run_id: 9629 and 9630) using the source-independent analysis.

For the analysis of GRB221009A for runs in moon conditions go to the following page: GRB221009A_src-ind_analysis_moon

The wiki page with the summary of the current results of the src-ind moon analysis (as of February 8th) for the discussion with LST-Reco conveners is the following: [1]

• Slides presented at the GRB221009A meetings about the source-independent dark/mild-moon analysis:

Start and Dark analysis I

Dark analysis II

Dark analysis III

Dark analysis IV and mild moon

Dark analysis V and mild moon

Dark analysis

The MC production for the Crab is used because the NSB tuning parameters at the DL1 level are very similar between them:

• 

  Pixel charge distribution for real cosmics (blue, run 9672), standard MC gammas (orange), tuned MC gammas to Crab (green) and to GRB221009A (red) using the DL1 NSB parameters. Plot with y-axis in log scale.

DL1 parameters obtained from GRB221009A runs (the NSB tuning parameters for the Crab can be found in the following subsection):

Runs: 9672, 9802, 9862 and 9890
{
    "number_of_files_used": 46,
    "std": [
    0.181,
    0.077,
    0.209
]
}{
    "increase_nsb": true,
    "extra_noise_in_dim_pixels": 1.782,
    "extra_bias_in_dim_pixels": 0.601,
    "transition_charge": 8,
    "extra_noise_in_bright_pixels": 2.25
}

Monte Carlo information

• Link to MC files used: /fefs/aswg/data/mc/DL1/AllSky/20221027_v0.9.9_crab_tuned/*
  • Particle types: The standard ones for the AllSky MC production.
  • DEC Band (deg): dec_2276

DL1 data

No reprocessing of the real DL1a data was required, only NSB tuning was added to the MC files. DL1a files produced by LSTOSA (lstchain v0.9.6?) and lstMCpipe (lstMCpipe v0.9.0 using lstchain v0.9.9).

• original DL1a files

/fefs/aswg/data/real/DL1/20221010/v0.9/tailcut84/dl1_LST-1.Run0XXXX.XXXX.h5
/fefs/aswg/data/mc/DL1/AllSky/20221027_v0.9.9_base_prod/

• lstchain v0.9.9
• real data: tailcut8-4 (with cleaning based on pedestal RMS), dynamic cleaning.

 /fefs/aswg/data/real/DL1/20221015/v0.9/tailcut84/log/lstchain_config_v0.9.4_from2022onwards.json

• lstMCpipe-v0.9.0
• lstchain v0.9.9
• MC: tailcut8-4, dynamic cleaning, NSB tuning

   "image_modifier": {
       "increase_nsb": true,
       "extra_noise_in_dim_pixels": 1.62,
       "extra_bias_in_dim_pixels": 0.655,
       "transition_charge": 8,
       "extra_noise_in_bright_pixels": 2.08,
       "increase_psf": false,
       "smeared_light_fraction": 0
   },

Config file: https://github.com/cta-observatory/lstmcpipe/blob/master/production_configs/20221027_v0.9.9_crab_tuned/lstchain_config_2022-10-27.json

Random forest

RFs were produced with lstMCpipe

• lstchain-v0.9.9

• lstMCpipe-v0.9.0

• source-indep (the standard one for the AllSky MC production)

/fefs/aswg/data/models/AllSky/20221027_v0.9.9_crab_tuned/dec_2276/

DL2 data

• lstchain-v0.9.9

• source-indep

/fefs/aswg/data/mc/DL2/AllSky/20221027_v0.9.9_crab_tuned/TestingDataset/dec_2276/node_theta_*
/fefs/aswg/workspace/arnau.aguasca/Analysis/results/real/DL2/202210*/srcind/v0.9.9/tailcut84_dynclg_tNSBtoCrabPerfPaper/AllSky_20221027_v0.9.9_crab_tuned_dec_2276/dl2_*

• config file

 /fefs/aswg/data/real/DL1/20221015/v0.9/tailcut84/log/lstchain_config_v0.9.4_from2022onwards.json

DL2 data selection to produce theta2 plots

Information about the quality cuts to produce the theta2 plots.

Example

• intensity > 50
• r: [0, 1 ]
• wl: [0.01, 1 ]
• leakage_intensity_width_2: [0, 0.2 ]
• event_type: [32, 32]

• source-indep
  • fixed_gh_cut: 0.7 (standard cut)
  • fixed_theta_cut: 0.2 (standard cut)

Theta2 plot

• 

  Theta2 plot using all runs in dark conditions. Three different energy bins are displayed

Crab check

The following Crab runs are used for the crab check: November 21 and 23 2022; 10914-8 and 10986-9. These runs are chosen because they were taken in similar atmospheric conditions and similar zenith and azimuth values.

Crab data is processed with the same analysis specifications as the GRB221009A runs in dark conditions.

• Quality cuts in gammaness and theta

Efficiency cuts in gammaness and theta are used. In particular, three gammaness efficiency values and one theta efficiency are used. The value of the theta efficiency is set to 70% following the performance paper. A higher value is not chosen to avoid reducing the signal-to-noise (since the first run in dark conditions started 5 days after the trigger).

Combinations:

• geff = 0.5 and teff = 0.7
• geff = 0.7 and teff = 0.7
• geff = 0.9 and teff = 0.7

The IRFs are produced using Chaitanya's branch with the nearest node option.

• The energy threshold for Crab runs:

The energy threshold is obtained using the closest node to the middle of the Crab runs distribution in cos(ZD)-sin(delta) plane.

• File:Ethd-Crab-GRB221009A.png

  MC true energy distribution to obtain the energy threshold.

• Spectral energy distribution:

The spectral energy distributions for the different cuts in gammaness and theta efficiency cuts show a discrepancy below 100 GeV.

• 

  SED using Crab runs with geff = 0.5 and teff 0.7. Lower energy bound for the fitting at E=50 GeV.
• 

  SED using Crab runs with geff = 0.7 and teff 0.7. Lower energy bound for the fitting at E=50 GeV.
• 

  SED using Crab runs with geff = 0.9 and teff 0.7. Lower energy bound for the fitting at E=50 GeV.

Thus, the energy threshold is increased to 100 GeV. The resulting SED is the following:

• 

  SED using Crab runs with geff = 0.5 and teff 0.7 (blue), geff = 0.7 and teff 0.7 (orange) and geff = 0.9 and teff 0.7 (green). Lower energy bound for the fitting at E=100 GeV.

The results are very similar between analyses. I decide to use the gammaness efficiency cut of 50% and an efficiency cut in theta of 70%.

Mild moon analysis

Due to the mild moon condition, the runs 9629 and 9630 show different NSB. Thus, the image cleaning may change w.r.t. the standard picture and boundary threshold.

• NSB

NSB tuning parameters for run_id = 9629:

{
"number_of_files_used": 16,
"std": [
    0.392,
    0.141,
    0.548
]
}{
    "increase_nsb": true,
    "extra_noise_in_dim_pixels": 3.145,
    "extra_bias_in_dim_pixels": 0.957,
    "transition_charge": 8,
    "extra_noise_in_bright_pixels": 3.995
}

NSB tuning parameters for run_id = 9630:

{
"number_of_files_used": 8,
"std": [
    1.176,
    0.276,
    1.816
]
}{
    "increase_nsb": true,
    "extra_noise_in_dim_pixels": 6.426,
    "extra_bias_in_dim_pixels": 1.656,
    "transition_charge": 8,
    "extra_noise_in_bright_pixels": 8.776
}

We can see that the NSB parameter values for 9629 and 9630 are x2 and x3 higher than the ones used for the dark conditions.

• Image cleaning

The picture threshold value is computed based on the condition that this value satisfies that only about 4% of the pedestal pixels will raise their picture thresholds due to the pedestal std dev condition, and not more than 5% of pedestals survive the cleaning

• 

  Picture value distribution (left) and its CDF (right) for run_id=9629. Vertical black line delimiters the 4% condition to set the picture threshold.
• 

  Picture value distribution (left) and its CDF (right) for run_id=9630. Vertical black line delimiters the 4% condition to set the picture threshold.

The picture threshold for run 9629 and 9630 is approximately 8 and 14, respectively. The boundary threshold value is set as half the picture threshold value.

• Analysis approach

In order to reduce the computational cost of the analysis for just two runs. I decided to make an MC production for 9630, but only tune the testing MC gammas for 9629 and produce IRFs with those MC files. Thus, the MC production 20221027_v0.9.9_crab_tuned is used for the 9629 run.

Monte Carlo information

• Link to MC files used: /fefs/aswg/data/mc/DL1/AllSky/20221027_v0.9.9_crab_tuned/* and /fefs/aswg/data/mc/DL1/AllSky/20221215_v0.9.12_base_prod/*
  • Particle types: The standard ones for the AllSky MC production.
  • DEC Band (deg): dec_2276

DL1 data

Reprocessing of the real DL1a data was required, only NSB tuning was added to the MC files. DL1a files produced by LSTOSA (lstchain v0.9.6?) and lstMCpipe (lstMCpipe v0.9.0 and lstMCpipe v0.10; using lstchain v0.9.9).

• original DL1a files

/fefs/aswg/data/real/DL1/20221010/v0.9/tailcut84/dl1_LST-1.Run0XXXX.XXXX.h5
/fefs/aswg/data/mc/DL1/AllSky/20221027_v0.9.9_crab_tuned/
/fefs/aswg/data/mc/DL1/AllSky/20221027_v0.9.9_base_prod/

• lstchain v0.9.9
• real data 9629: tailcut8-4 (with cleaning based on pedestal RMS), dynamic cleaning.

 /fefs/aswg/data/real/DL1/20221015/v0.9/tailcut84/log/lstchain_config_v0.9.4_from2022onwards.json

• lstMCpipe-v0.9.0
• lstchain v0.9.9
• MC training: tailcut8-4, dynamic cleaning, NSB tuning

   "image_modifier": {
       "increase_nsb": true,
       "extra_noise_in_dim_pixels": 1.62,
       "extra_bias_in_dim_pixels": 0.655,
       "transition_charge": 8,
       "extra_noise_in_bright_pixels": 2.08,
       "increase_psf": false,
       "smeared_light_fraction": 0
   },

Config file: https://github.com/cta-observatory/lstmcpipe/blob/master/production_configs/20221027_v0.9.9_crab_tuned/lstchain_config_2022-10-27.json

• MC testing: tailcut8-4, dynamic cleaning, NSB tuning

   "image_modifier": {
       "increase_nsb": true,
       "extra_noise_in_dim_pixels": 1.62,
       "extra_bias_in_dim_pixels": 0.655,
       "transition_charge": 8,
       "extra_noise_in_bright_pixels": 2.08,
       "increase_psf": false,
       "smeared_light_fraction": 0
   },

Config file: /fefs/aswg/workspace/arnau.aguasca/scripts/_configs/GRB221009A/lstchain_MC_config_lstchain_belong_v0.9.9_NSBDL1to9629.json

• real data 9630: tailcut14-7 (with cleaning based on pedestal RMS), dynamic cleaning.

 /fefs/aswg/workspace/arnau.aguasca/scripts/_configs/GRB221009A/real_config_lstchain_belong_v0.9.9_tailcut147_9630.json

• lstchain v0.9.9
• MC: tailcut14-7, dynamic cleaning, NSB tuning

   "image_modifier": {
       "increase_nsb": true,
       "extra_noise_in_dim_pixels": 6.426,
       "extra_bias_in_dim_pixels": 1.656,
       "transition_charge": 8,
       "extra_noise_in_bright_pixels": 8.776
       "increase_psf": false,
       "smeared_light_fraction": 0
   },

Config file: /fefs/aswg/workspace/arnau.aguasca/scripts/_configs/GRB221009A/lstchain_MC_config_lstchain_belong_v0.9.9_tailcut147_NSBDL1to9630.json

Random forest

RFs were produced with lstMCpipe

• lstchain-v0.9.9

• lstMCpipe-v0.9.0 (for run 9629) and lstMCpipe-v0.9.0 (for run 9630)

• source-indep (the standard one for the AllSky MC production)

/fefs/aswg/data/models/AllSky/20221027_v0.9.9_crab_tuned/dec_2276/
/fefs/aswg/workspace/MC_data_simlink/models/AllSky/20230123_GRB221009A_moon_mild_ph/dec_2276/

DL2 data

• lstchain-v0.9.9

• source-indep

/fefs/aswg/data/mc/DL2/AllSky/20221027_v0.9.9_crab_tuned/TestingDataset/dec_2276/node_theta_*
/fefs/aswg/workspace/MC_data_simlink/DL2/AllSky/20230123_GRB221009A_moon_mild_ph/TestingDataset/dec_2276/node_theta_*  
/fefs/aswg/workspace/arnau.aguasca/Analysis/results/real/DL2/20221015/srcind/v0.9.9/tailcut84_dynclg_tNSBtoCrabPerfPaper/AllSky_20221027_v0.9.9_crab_tuned_dec_2276/dl2_*
/fefs/aswg/workspace/arnau.aguasca/Analysis/results/real/DL2/20221015/srcind/v0.9.9/tailcut147_dynclg_tNSBmildMoon9630/AllSky_20230123_GRB221009A_moon_mild_ph_dec_2276/dl2_*

• config file

 /fefs/aswg/data/real/DL1/20221015/v0.9/tailcut84/log/lstchain_config_v0.9.4_from2022onwards.json
 /fefs/aswg/workspace/arnau.aguasca/scripts/_configs/GRB221009A/real_config_lstchain_belong_v0.9.9_tailcut147_9630.json
 /fefs/aswg/data/real/DL1/20221015/v0.9/tailcut84/log/lstchain_config_v0.9.4_from2022onwards.json
 /fefs/aswg/workspace/arnau.aguasca/scripts/_configs/GRB221009A/lstchain_MC_config_lstchain_belong_v0.9.9_NSBDL1to9629.json

DL2 data selection to produce theta2 plots

Information about the quality cuts to produce the theta2 plots.

Example

• intensity > 50
• r: [0, 1 ]
• wl: [0.01, 1 ]
• leakage_intensity_width_2: [0, 0.2 ]
• event_type: [32, 32]

• source-indep
  • fixed_gh_cut: 0.7 (standard cut)
  • fixed_theta_cut: 0.2 (standard cut)

Theta2 plot

• 

  Theta2 plot using run_id = 9629 and 9630.

Analysis Results

The IRFs are produced using Chaitanya's branch with the nearest node option.

The high-level analysis is done by dividing the runs into three different datasets based on:

• The more distant to T0, the more runs together
• Runs taken at similar dates
• Significance

The datasets are divided as follows:

• subset 1: 2022-10-15
• subset 2: 2022-10-[16-18]
• subset 3: 2022-10-[21-28]

Spectral results

• We use the quality cut in efficiency gammaness and theta of 50% and 70%, respectively (see Crab check).

• The energy threshold is set to 100 GeV (see Crab check)

• The assumed spectral model is set to a power law with a fixed spectral index of -2. The contribution of the EBL is considered using the available redshift information. A joint fitting is considered.

• 

  GRB221009A SED for subset 1 with geff = 0.5 and teff 0.7. Lower energy bound for the fitting at E=100 GeV.
• 

  GRB221009A SED for subset 2 with geff = 0.5 and teff 0.7. Lower energy bound for the fitting at E=100 GeV.
• 

  Fit residuals for subset 3 with geff = 0.5 and teff 0.7. Lower energy bound for the fitting at E=100 GeV.

The fit converged for the subset 3, but the predicted signal counts do not match the excess counts.

Light curve

• The minimum energy to compute the integral flux is set to two times the energy threshold (emin=200 GeV) in order to reduce the systematics.

• 

  Light curve. Lower energy bound to compute the integral flux at E=200 GeV.