BLLac paper meeting

From my_wiki
Revision as of 17:04, 19 January 2023 by Nozaki (talk | contribs) (→‎collaborator meeting)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

collaborator meeting[edit]

  • 2023-01-17
    • general remarks (Seiya) slides
      • Finer IRF interpolation: Chaintanya checked the distribution of Az/Zd during the run and confirmed there is no much difference. So the interpolation using the single pointing information for each run is enough or has a minor effect than other systematics.
      • optical and IR data can be used from the published paper
      • ICRC contribution by Seiya is approved within the group.
      • Let's try to get the preliminary final plots during the spring, then the target date for the paper submission is within this year (or ealier).
      • overleaf link is distributed to put the preliminary results
    • Likelihood analysis (Gabriel) slides
      • Update on likelihood analysis using source-dependent approach
      • On Aug 8, the averaged spectrum looks almost consistent with Chaitanya's results, but there is some discrepancy on the light curve. We also saw such kind of discrepancy between different RF with different nodes (dec3476 vs merged) for source-indep analysis. Gabriel try to check using another nodes for RF training, and see if we have similar trends.
      • On Aug 3, there is discrepancy even on SED. Need to be checked.
      • For the final analysis, we need to check the systematics using different efficiency cuts as the performance paper.
    • Standard analysis (Chaitanya) slides
      • Using two off regions gives us higher significance, and we also get some additional data points above TeV.
      • If the selected regions for ON and OFF are overlapped, it should be excluded. Does it work also for energy-dependent theta cut? - Yes, it should work.
    • SED modeling (Juan) slides
      • Trying to model the MWL SED. First trial using the same parameters as Prince paper.
      • Best model so far assumes UV data as the thermal emission. But HE does not fit well.
      • First, let's try to model using a simple SSC.
    • SED evolution (Seiya) MWL daily SED Aug2 Aug3 Aug8 Aug9 LC
      • intra-night SED evolution using a single/two runs corresponding to the flaring state.
      • The assumed model shape is very different between Aug 4 and Aug5 -> can be explained by the different of zenith angles (low Zd vs high Zd)
      • To evaluate the spectral change quantitatively, need to fix the energy range


  • 2022-11-01
    • update (MWL analysis) slides
    • MWL LC
      • LST-1 + Fermi/LAT + Swift/XRT + Swift/UVOT
      • there are two components with different time scale (day-scale variability + hour-scale variability)
      • There are Swift data around the peak of Aug 3, but no much data around the peak of Aug 8
      • it is a pity we don't have Swift data around Aug 8, but can we expect the evolution of other wavelength data (UV & X-ray) based on the existing data (gamma)? - It would be difficult since the timescale of the emission is different (= those emissions are not necessarily based on the same component.)
    • MWL SED
      • LST spectrum is observed one or intrinsic one? - the plotted data are observed one. For the modeling, let's use the intrinsic spectrum (model curve) and not consider EBL absorption in the modeling
      • X-ray spectrum shape is a bit strange, the absorption was already corrected? - To be checked, but it was the average spectrum for 10 days, so this might be the reason of the shape (Seiya)
      • basically the peak of the MWL spectrum is not much different from 2015 flare. Just a flux level is much higher than before. It will need a lower magnetic field and etc..
      • MWL modeling will be tried by Juan and Katsuaki. The modeling should be done by the standard tool (jetset).
      • how about the evolution of SED (day-wise or run-wise)? - For the LST analysis, if the flux level is low, we may need to stack the dataset, but can be done. For MWL SED, Seiya will produce day-wise MWL SED soon, then let's consider again how to proceed for the final analysis.
    • Request for a new MC production
      • need to request a new MC production using the latest version of lstchain (NSB tuning was fixed)
      • simple "tel_az" should be fine instead of sin_tel_az
        • If we use north-pointing data with Crab-path-RF, this az-issue could be problematic. But if we train RF using north-pointing MC for north-pointing data, it should not be a problem. Better not to use a function of some parameters (Chaitanya)
        • anyway we cannot use sin_tel_az using the released version of lstchain
      • Let's request merged-node production (dec 3476 + 4622)
      • In a configuration file, we can activate a flag to save parameters for likelihood analysis. So let's activate it for the MC production (Gabriel). Seiya will prepare a configuration file for production and share it within the group to confirm if this flag is properly activated.
    • LST analysis
      • which SED can we show in the paper? - the possible idea is 1) Aug 3, 2) Aug 8, 3) low state.
      • concerning the intra-night variability, which night can we show in the paper? - it is a pity that theare is a long (~2 hours?) break between the last runs and others... but why not, we can show it, so the possible idea is two highest nights (Aug 3 and Aug 8)
      • it is also interesting to show the evolution of the index of the spectrum
      • how about Fermi+LST joing fitting? - it is also interesting, but there is a bug in gammapy for the joing fit using Fermi (3D) and LST (1D). After the bugfix, let's try.
    • Target data
      • Let's have preliminary results for all the analysis within this year...!
        • Seiya will be in the MAGIC shift, and Gabriel will move to Marseille (congratulations!)



  • 2022-05-12
    • short update slides
      • Our analysis have almost consistent results, though there are small (not large) discrepancy. It should be verified later with the final MC.
      • All sky MC production is ongoing. One of dec_3476 was already partially ready on DL1/2 level.
        • A new standard configuration can make the time of RF training shorter than before (1day -> 2 hours)
        • a single RF using alt/az or separate of data set with corresponding poiting RF: which analysis can improve the performance? Need to check.
      • NSB level of BLLac FoV is slightly higher than other egal source due to low galactic latitude. Need to check the difference on parameter distribution with/without NSB tuning. If the effect is not negligible, we need to send a request of NSB tuning also for BLLac.
        • NSB tuning also can be added in R0 level


  • 2022-04-08
    • Statu update
      • (Seiya) checking the data quality also based on intensity distribution. Flux level in LC is a bit lower than the previous analysis slides (Seiya)
      • (Chaitanya) good observation runs are analyzed with gamma/theta efficiency cut (without intensity cut). SED on the night of flare is consistent with the Seiya's previous analysis slides (Chaitanya)
      • (Gabriel) LHfit code implementation is ongoing. Crab spectrum looks more or less fine. BLLac spectrum below ~80GeV is higher than the previous analysis. Need to be checked. slides (Gabriel)
      • (Juan) Flux level in LC is a bit lower than the previous analysis. More or similar as what Seiya currently obtained. To be investigated. (Juan)
    • Presentation from BLLac LIV team by Cyann
      • analysis of the light curve on the night of flare with different energy ranges (<~200GeV, >~200GeV). LC of the night of Aug 9 (next day of the brightest flare) will be also checked by our team
      • we keep the current scheme (analysis independently), then we provide the high-level analysis product (DL3) with LIV team at some moment for the publication.
    • Juan's presentation@SEA Spanish national meeting
    • Next step for the LST general meeting
      • Let's make a comparison plot with the same analysis conditions (basic shower parameter cut, energy range/binning for SED, LC). The detailed conditions will be announced
    • Next MC production
      • All sky MC for BLLac path will be ready in two weeks or so. But we also need IRF interpolation software. Anyway let's discuss later again how to proceed after the MC dataset for BLLac will be prepared.


  • 2022-03-10 slides
    • We are waiting for the MC grid production for this source, and also need to wait for interpolate-IRF software
    • We will try to produce results with existing MC/data with the latest version of lstchain before the next meeting (in two weeks), then we can compare those results.
    • For the selling point of this paper, we will continue to discuss with analysis results which will be produced and get comments in extragalactic meeting.
    • lst1-bllac-paper slack channel is created
Retrieved from "https://www.lst1.iac.es/wiki/index.php?title=BLLac_paper_meeting&oldid=6754"