Difference between revisions of "Analysis A. Aguasca-Cabot lstchain v0.9.X AllSky"
(One intermediate revision by the same user not shown) | |||
Line 1: | Line 1: | ||
= Overview = | = Overview = | ||
*Source-dependent/independent analysis using LSTCHAIN v0.9.6 and the branch "interp_irfs" with last commit e5835d5ab7792cbb528118de9cebe4a60f7802a8. The MC production processed with LSTOSA v0.8.2 using LSTCHAIN v0.9.6 | *Source-dependent/independent analysis using LSTCHAIN v0.9.6 and the branch "interp_irfs" with last commit e5835d5ab7792cbb528118de9cebe4a60f7802a8. The MC production processed with LSTOSA v0.8.2 using LSTCHAIN v0.9.6 | ||
− | *Cleaning method tail cut 84 and dynamic cleaning. Also, the MC files are tuned to match the NSB in the FoV | + | *Cleaning method tail cut 84 and dynamic cleaning. Also, the MC files are tuned to match the NSB in the FoV. |
*Observations used in the analysis: | *Observations used in the analysis: | ||
Line 11: | Line 11: | ||
:- Dec band (deg): dec_min_431 | :- Dec band (deg): dec_min_431 | ||
:- Other information: MC production through PR in lstMCpipe Github webpage (https://github.com/cta-observatory/lstmcpipe/tree/master/production_configs/20220523_dec_413_tuned_nsb) | :- Other information: MC production through PR in lstMCpipe Github webpage (https://github.com/cta-observatory/lstmcpipe/tree/master/production_configs/20220523_dec_413_tuned_nsb) | ||
− | :- MC files are tuned to match the NSB in the FoV | + | :- MC files are tuned to match the NSB in the FoV. |
== DL1 data == | == DL1 data == | ||
Line 51: | Line 51: | ||
− | |||
− | |||
− | |||
− | |||
− | |||
* Original DL1a files | * Original DL1a files | ||
/fefs/aswg/data/mc/DL1/AllSky/galsource_min_413_tuned_nsb/{TrainingDataset,TestingDataset}/dec_min_413/{particle}/node... | /fefs/aswg/data/mc/DL1/AllSky/galsource_min_413_tuned_nsb/{TrainingDataset,TestingDataset}/dec_min_413/{particle}/node... | ||
Line 68: | Line 63: | ||
/fefs/aswg/workspace/arnau.aguasca/scripts/_configs/lstchain_src_dep_tailcut84_config_dispnorm_v0.9.4.json | /fefs/aswg/workspace/arnau.aguasca/scripts/_configs/lstchain_src_dep_tailcut84_config_dispnorm_v0.9.4.json | ||
:- path: | :- path: | ||
− | /fefs/aswg/workspace/MC_data_simlink/models/AllSky/ | + | /fefs/aswg/workspace/MC_data_simlink/models/AllSky/galsource_min_413_tuned_nsb_NOpsf_srcdep/dec_min_413/ |
*; Source-independent | *; Source-independent | ||
:- Processed with lstMCpipe using LSTCHAIN v0.9.6. | :- Processed with lstMCpipe using LSTCHAIN v0.9.6. | ||
:- Path: | :- Path: | ||
− | /fefs/aswg/ | + | /fefs/aswg/workspace/MC_data_simlink/models/AllSky/galsource_min_413_tuned_nsb_NOpsf/dec_min_413/ |
+ | |||
+ | :: Notice that the path moved from aswg/data to aswg/workspace/MC_data_simlink. The MC prod in aswg/data has a smeared PSF (PSF tuning). | ||
== DL1 to DL2 data == | == DL1 to DL2 data == | ||
Line 101: | Line 98: | ||
== Quality cuts optimization == | == Quality cuts optimization == | ||
− | Since the AllSky MC is a tailored production than the Fixed MC production, we cannot use high-zenith Crab observations to find the best cuts. Thus, we have to use MC data to find the best cuts for RS Oph. | + | Since the AllSky MC is a tailored production than the Fixed MC production, we cannot use high-zenith Crab observations to find the best cuts. Thus, we have to use MC data to find the best cuts for RS Oph. I did the following: |
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
*Procedure | *Procedure | ||
− | - | + | :1- I weighted the Testing DL2 MC dataset to the spectral index of RS Oph without considering RS Oph amplitude, i.e. maximizing the surviving MC gammas :following a spectra index of -4. |
− | - | + | :2- Cut in theta (< 0.3 deg) and apply filter (remove the contribution of events due to the long tails of the PSF) |
− | + | :3- Obtain the gammaness cut for a percentile of 90% survival gammas for each energy bin: Energy array with 5 bins per decade | |
− | - | + | :4- Apply the energy dependent cuts in gammaness to the unweighted testing MC DL2 files to obtain the IRFs |
− | |||
− | |||
== DL3 data selection == | == DL3 data selection == | ||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
*:- Quality cuts | *:- Quality cuts | ||
Line 148: | Line 116: | ||
"length": [0, Infinity], | "length": [0, Infinity], | ||
"r": [0, 1], | "r": [0, 1], | ||
− | "wl": [0. | + | "wl": [0.01, 1], |
"leakage_intensity_width_2": [0, 0.2], | "leakage_intensity_width_2": [0, 0.2], | ||
"event_type" : [32, 32] | "event_type" : [32, 32] | ||
Line 154: | Line 122: | ||
}, | }, | ||
"DL3FixedCuts": { | "DL3FixedCuts": { | ||
− | "fixed_gh_cut": 0. | + | "fixed_gh_cut": 0.6, |
− | "fixed_theta_cut": 0. | + | "gh_efficiency": 0.9, |
+ | "fixed_theta_cut": 0.2, | ||
+ | "theta_containment": 0.68, | ||
+ | "alpha_containment": 0.68, | ||
"allowed_tels": [1] | "allowed_tels": [1] | ||
} | } | ||
Line 162: | Line 133: | ||
*; Source-dependent | *; Source-dependent | ||
*:- IRF: point-like, single-offset | *:- IRF: point-like, single-offset | ||
− | /fefs/aswg/workspace/arnau.aguasca/scripts/_configs/ | + | /fefs/aswg/workspace/arnau.aguasca/scripts/_configs/DL3IRF-generated-configs/config_int50_leak0.2_gh0.7_geff0.9_th0.2_tc0.68_a15_ac0.68.json |
::- Produced DL3 files | ::- Produced DL3 files | ||
− | /fefs/aswg/workspace/arnau.aguasca/ | + | /fefs/aswg/workspace/arnau.aguasca/Analysis/results/real/DL3/RSOph/nearest_node/... |
*; Source-independent | *; Source-independent |
Latest revision as of 13:37, 16 September 2022
Contents
Overview[edit]
- Source-dependent/independent analysis using LSTCHAIN v0.9.6 and the branch "interp_irfs" with last commit e5835d5ab7792cbb528118de9cebe4a60f7802a8. The MC production processed with LSTOSA v0.8.2 using LSTCHAIN v0.9.6
- Cleaning method tail cut 84 and dynamic cleaning. Also, the MC files are tuned to match the NSB in the FoV.
- Observations used in the analysis:
Monte Carlo information[edit]
- AllSky MC prodution.
- Link to MC files used: /fefs/aswg/data/mc/DL1/AllSky/galsource_min_413_tuned_nsb/...
- - Particle types: gamma diffuse and protons
- - MC prod (deg): galsource_min_413_tuned_nsb
- - Dec band (deg): dec_min_431
- - Other information: MC production through PR in lstMCpipe Github webpage (https://github.com/cta-observatory/lstmcpipe/tree/master/production_configs/20220523_dec_413_tuned_nsb)
- - MC files are tuned to match the NSB in the FoV.
DL1 data[edit]
- Processed with LSTOSA using LSTCHAIN v0.9.6.
- - Dynamic cleaning and tail cut cleaning with pedestal threshold applied by the the standard parameters using pipeline LSTOSA.
"dynamic_cleaning": { "apply": true, "threshold": 267, "fraction_cleaning_intensity": 0.03 }
Real data[edit]
- Parameters for the tail cut cleaning with pedestal threshold
"tailcuts_clean_with_pedestal_threshold": { "picture_thresh":8, "boundary_thresh":4, "sigma":2.5, "keep_isolated_pixels":false, "min_number_picture_neighbors":2, "use_only_main_island":false, "delta_time": 2 }
- Original DL1a files (processed up to DL1 by LSTOSA using LSTCHAIN v0.9.2)
/fefs/aswg/data/real/DL1/{}/v0.9/tailcut84/dl1_LST-1.RunXXXXX.XXXX.h5
MC data[edit]
- Tuned NSB
"increase_nsb": true, "extra_noise_in_dim_pixels": 0.937, "extra_bias_in_dim_pixels": 0.323, "transition_charge": 8, "extra_noise_in_bright_pixels": 1.041,
- Original DL1a files
/fefs/aswg/data/mc/DL1/AllSky/galsource_min_413_tuned_nsb/{TrainingDataset,TestingDataset}/dec_min_413/{particle}/node...
Random forest[edit]
- Dispnorm parameterisation
- Source-dependent
- - LSTCHAIN v0.9.6
- - Config file:
/fefs/aswg/workspace/arnau.aguasca/scripts/_configs/lstchain_src_dep_tailcut84_config_dispnorm_v0.9.4.json
- - path:
/fefs/aswg/workspace/MC_data_simlink/models/AllSky/galsource_min_413_tuned_nsb_NOpsf_srcdep/dec_min_413/
- Source-independent
- - Processed with lstMCpipe using LSTCHAIN v0.9.6.
- - Path:
/fefs/aswg/workspace/MC_data_simlink/models/AllSky/galsource_min_413_tuned_nsb_NOpsf/dec_min_413/
- Notice that the path moved from aswg/data to aswg/workspace/MC_data_simlink. The MC prod in aswg/data has a smeared PSF (PSF tuning).
DL1 to DL2 data[edit]
Real data[edit]
- Source-dependent
- - LSTCHAIN v0.9.6
- - Config file:
/fefs/aswg/workspace/arnau.aguasca/scripts/_configs/RSOph/analysis_v0.9.6/lstchain_config_srcdep_lstmcpipe_tailcut84_dispnorm_v0.9.6.json
- Source-independent
- - LSTCHAIN v0.9.6
- - Config file:
/fefs/aswg/workspace/arnau.aguasca/scripts/_configs/RSOph/analysis_v0.9.6/lstchain_config_lstmcpipe_tailcut84_dispnorm_v0.9.6.json
MC data[edit]
- Source-dependent
- - Processed with lstMCpipe using LSTCHAIN v0.9.6.
- - Config file:
/fefs/aswg/workspace/arnau.aguasca/scripts/_configs/RSOph/analysis_v0.9.4/lstchain_config_srcdep_lstmcpipe_tailcut84_dispnorm_v0.9.6.json
- Source-independent
- - Processed with lstMCpipe using LSTCHAIN v0.9.6.
- - Config file:
/fefs/aswg/data/models/AllSky/galsource_min_413_tuned_nsb/dec_min_413/lstchain_config_2022-05-23.json
Quality cuts optimization[edit]
Since the AllSky MC is a tailored production than the Fixed MC production, we cannot use high-zenith Crab observations to find the best cuts. Thus, we have to use MC data to find the best cuts for RS Oph. I did the following:
- Procedure
- 1- I weighted the Testing DL2 MC dataset to the spectral index of RS Oph without considering RS Oph amplitude, i.e. maximizing the surviving MC gammas :following a spectra index of -4.
- 2- Cut in theta (< 0.3 deg) and apply filter (remove the contribution of events due to the long tails of the PSF)
- 3- Obtain the gammaness cut for a percentile of 90% survival gammas for each energy bin: Energy array with 5 bins per decade
- 4- Apply the energy dependent cuts in gammaness to the unweighted testing MC DL2 files to obtain the IRFs
DL3 data selection[edit]
- - Quality cuts
"EventSelector": { "filters": { "intensity": [50, Infinity], "width": [0, Infinity], "length": [0, Infinity], "r": [0, 1], "wl": [0.01, 1], "leakage_intensity_width_2": [0, 0.2], "event_type" : [32, 32] } }, "DL3FixedCuts": { "fixed_gh_cut": 0.6, "gh_efficiency": 0.9, "fixed_theta_cut": 0.2, "theta_containment": 0.68, "alpha_containment": 0.68, "allowed_tels": [1] }
High-level analysis[edit]
- Source-dependent
- - IRF: point-like, single-offset
/fefs/aswg/workspace/arnau.aguasca/scripts/_configs/DL3IRF-generated-configs/config_int50_leak0.2_gh0.7_geff0.9_th0.2_tc0.68_a15_ac0.68.json
- - Produced DL3 files
/fefs/aswg/workspace/arnau.aguasca/Analysis/results/real/DL3/RSOph/nearest_node/...
- Source-independent
Analysis results[edit]
- High level analysis performed with gammapy-v0.20.1
Alpha plot[edit]
- Filters in EventSelector applied:
"EventSelector": { "filters": { "intensity": [50, Infinity], "width": [0, Infinity], "length": [0, Infinity], "r": [0, 1], "wl": [0.1, 1], "leakage_intensity_width_2": [0, 0.2] } }
See section Quality cuts optimization to know more about the justification of the applied cuts.