Notes on eSASS telecon of 2017-10-16

Participants: GL, AG, IK, TD, JS, CG, MF, HB, KD, first time participants: Mara Salvato, Teng Liu

Teng Liu would like to get involved in eSASS work, specifically in the area of source detecion. We also discussed his potential involvement in the area of astrometric corrections of the source catalogs.

eROSE (TD): first task was building low level framework. Current functionallity: limited catalog browsing, accessing public ancilliary data (image, overlays), now working on functionality to allow users to upload their own catalogs. Detailed data model for eROSITA products still needs to be built.

MS: its important to document properties/limitation of the catalogs (covered area, flux limit, etc.).

HB: internal alpha test version of eROSE would be useful.

Catalog cross-matching (MS): plans for follow-up/identification meeting. Prior for eROSITA depths (shallow and deep fields) needs to be defined. This is different from ROSAT or XMM work, WISE prior that was used is not suitable for eROSITA. Now working on Stripe 82X and 3XMM-DR7 data, defining optimal bands for prior (machine learing). It is not clear yet whether the X-ray flux will be part of the prior.

Mara's multi-lambda source identification scripts will run independently of eSASS. The function of eROSE in the multi-lambda work is not clear yet (depends on whether X-ray fluxes will be part of the prior). The multi-lambda pipeline assigns probabilities of sources being the correct counterpart of eROSITA sources. These probabilities will be fed into eROSE as an enhancement of the X-ray catalogs.

Preprocessor (IK): Review of recent changes of the tmsplit output. Ingo already sent (in September) example tmspit fits output files containing the ENV format columns and an extension (named CEstatus) containing the status words. The ENV format is currently still evolving somewhat, presumably requiring further adjustments in tmsplit. But the fits event file format seems to be ok as currently defined. It is also agreed that the on-board software will not split a 64 bit event (32 bit trigger word + 32 bit environment word) across different telemetry frames.

Ingo asked Diogo about the byte order of the time tags. Sigi has new Radiocomplex test datasets, which may be suitable for Ingo. These don't have proper time tags yet, these are only expected early next year (Michael). The headers may also not be 100% correct.

Byte order of the mock attitude data provided by Igor: This is a longstanding open question. Will need to ask Igor once again (done: he believes they are correct, will need to be confirmed with ground-test data).

Joint preprocessor/eSASS TEL chain test: Ingo first needs to update his SIXTE to raw telemetry converter to the current formats. Conversion of attitude to startracker telemetry also needs to be updated.

Plans to place vignetting in EV_WEIGTH column of event table and use non-vignetted exposure maps in ermldet. This would be a new option in ermldet to work with event based detection.

Flare screening (JS): speed-up by removing unnecessary loop iterations, change from PI to ENERGY column, talked to Nicolas and Stephania, moving from global flare GTI to to flare GTI per camera, but statistics of all eROSTIA necessary for time resolved flares.

SRCTOOL expects flare GTI per camarea, EVTOOL expects global flare GTI. We agreed that all GTI should be per camera. Needs to be changed in FLAREGTI and EVTOOL.

Source detection (AG, GL): reviewing work on shapelet PSF (modeling width of PSF core, integration over pixels, third scale for core, Konrad's up-sampling, understanding positional offset, possible SIXTE issues: PSFs, background scaling/non-Poissonian background?) - details discussed in dedicated telecons of AG, GL and HB.

Sensitivity maps: APETOOL will both add columns to ERMLDET source table AND create aperture photometry sensitivity maps (AG: this is easier than doing this in two separate tools). ERSENSMAP will create Maximum Likelihood sensitivity maps. ERMLDET already writes (empty) APETOOL columns to source lists, AG: is naming convention for new APETOOL columns ok?

Source list columns: GL implemented OWNER column (same algorithm as for event flag but written as one Integer number instead of two bit-flags). Source has IKI/MPE ownership if all photons in extraction area have IKI/MPE ownership.

CATPREP (CG): Splitting up CATPREP into four sub-tasks, off-line discussions by CG+HB, current plans in brief:

Testing SRCTOOL lightcurves.

Barycentric corrections (IK): initial activities in Bamberg: FTOOL based with wrapper for SRG orbit contribution. Ask IKI whether orbit file format in ICD is still current.

Discussion of astrometric corrections - possible approach:

EROSITAwiki: EroCat/eSASSdevel/telCoNotes/2017-10-16 (last edited 2017-11-17 13:45:20 by HermannBrunner)