Yan Grange : ARTS Apertif : transform the WSRT into an efficient 21 cm survey instrument useful also for Transient survey ARTS real time system Survey 2019 Filled priorities (a) Known fields for localization + new detections (b) Known fields Data products (1FRB per 7 days) DR1 PSRFITS Detections on disk non detections on tape two cone search FRB detection cone search FRB observation cone search + FRB properties catalog some data are in TNS, some in DACHS Maybe generate a TAP on the fly using the external info Distinguish between detections and non-detections ObsCore : all data / only those on disk ? Baptiste question : Is there any chance to have solar system observation ? Mark Kettenis : having on tape ? is too difficult. Good to know that detections are on disk . VG: Regarding how many missed FRBs may reside on tape libraries wordwide, there is an interesting research note here: https://iopscience.iop.org/article/10.3847/2515-5172/ab0a04 FB : most of the paraemeteres you have in the SCS can be mapped in ObsCore ------------------------------------- Alessandra : Mapping INAF single dish data into ObsCore 3 sites in Italy 3 observing modes SD, Pulsar, VLBI These are raw data . Single dish data 4 modes On/off scan, raster map scan, on the fly cross scan, on the fly map scan + additional frequency switching skydip (for calibration) Single dish data format : directory per scan Summmary fits files plus file per sub scan Mapping SD data into Obscore many attributes OK ... dataproduct_type (spectrum cube according to modes) OTF cross scan, skydip scan. Spatial profile forv OTF cross scan ? more than 1 spectral window in the same scan zoom mode (increasing resolution) different for each polarization dataproduct_type : multifrequency ? s_fov, s_region, s_xel1, s_xel2 multi feed receiver = several circles ... or bands ---> How to describe s-region ? complex polygon ? t_resolution = NULL delta_t is something else (because each time slot is meausring a different point in space) em_res_power : we prefer em_resolution (in units of frequency) em_res_power comes from optical and varies too much Q: do you need to distinguish between the two OTF scan modes? Answer : one is spatial profile, the other one is a cube Multifreq / Product types James : do you use dataproduct_subtype ? free text ? AZ: we considered that MM : Free text -- not controlled, might loose interoperability JD : @Marco - yes that’s the trade-off PT : Is the multi-freq problem not solved? ALMA does it all the time. FB : why not using one dataset for each spectral window/ resolution instead of type = multifrequency ? JD : For ATCA data we use a row per frequency window in obscore, but that of course means multiple rows per data file VG : I think it's a matter of trade-off in order not to puzzle who makes queries with lots of ObsCore records for each observation, as it may occurr for spectral line observations, while for continuum with say 4 spectral windows, use a row for each may be fine. As far as I know, this should be the ALMA case freq resolution Alberto : in Obscore resolution : we are not using resolution but resolution power / --> why we did that ? There was no unit issue MD : I've not quite got why you cannot give em_res_power? delta_lambda/lambda is nicely computable for your data, too, no? MK : delta_lambda varies across the band (since delte_freq is a fixed number). Not a real problem for small bandwidths but if y=the fractional bandwidth is large it could be a concern. AM : But it is not true that delta lambda / lambda == delta freq / freq ? MK : mathematically that is true ;) MD : Of course the resolving power isn't constant in optical spectra either. Mark, are you saying it's particularly nasty in radio because your spectra are so wide? Frankly, I'd not worry about specrp varying by a small factor over a spectrum. It's just a rough characterisation only anyway. MK : yes, for example consider a telescope observing from 80-400 MHz with a fixed resolution of 100 kHz MD : Ah, that's a factor of five... I'd be fine with that. A mean or even the worst value would, I claim, be about right for obscore. But I give you 5 octaves are a mouthful. But that´s a client issue, no? MK : yes BC : One option (we have implemented in EPNcore): is to use a spectral sampling keyword (in Hz) and a resolution in relative units (resolving power) Using MOC to describe s_region MK : Use MOC for s_region (as has been suggested earlier in this interim IIRC)? MD : Yeah, that's what I'd suggest with shapes like those. ---------------------------- FB : Extension Note to become a WD ? Remaining issues : frequency bounds --> official UDF or view with officila attributes uv plane description uv_distance_min and max = how to compute ? does the extension fit opticla interferometry ? implementation . Two tables join on obs_publisher_did add aseco,nd MODEL element in the capability + a description of the table in the tableset Implementation review to be ugraded at next interop. May we use it in the Springer book "data intesnive radio astronomy" where many of us are involved ? MD : please don't use TAPRegExt any more. Use table utypes. (as in EPN-TAP, LineTAP, RegTAP 1.1...) MD : TAPRegExt model was a mistake, as "model" is a property of schemas or tables, not of services. MM : @Markus: I suppose this needs be fixed in ObsTAP too having a common way to handle those extensions would be needed (e.g. refkeys to the main obscore) BC : Alan Loh to contribute to the extension note François; did you get feedback from NenuFAR guys ? As Alan presented in the previous Radio-TDIG session, there will be an ObsTAP service with NeunFAR data. So they have to review the note… VG (to FB) : Regarding fractional bandwidth, for radio interferometry (especially for imaging) it may be actually of interest, as higher is that ratio, more it degrades sensitivity and resolution moving further from the phase centre ___________________________________________ JIVE ObsTAP service for EVN dVLBI ERIC EVN : network JIVE instyiytue FITS-finder 15 years old content : visibility data / not images Too much possible reesolutions . (imaging depend on scientific goal) VO use case 1 Access historic data (GRB, Gravitational waves, FRB) 2 standardized access to archive data for science platform Jupiter Lab environment VO : ObsCore representation of visibility data multiplicity obs_id from the same observation obs_publisher_did each Obscore dataset 20K Obscore datasets / 2000 observations [~ 10 datasets per observation) DataLink calibration data is now added / amplitude calibration flagging information UVFLG Returned as calibration rows implementaion DacHS 2.5 EuroVO registration didn't propagate to other registries using Dachs istelf for registration worked Plot on Aladin spatial coverage can we use STMOC ??? does that make sense ? (too sparsed ?) Markus : use Dachs for coverage ? François : calibrated visibilities ? No we provide raw visibilities and calibration material to be used. Markus : single dish spectra ? JD: CASDA has an SSA service and it should have data these days - I’ll check on that MD : James: can you point me to the registry record of your SSA service? JD : ivo://au.csiro/casda/SSA MD : James: that service says its waveband is UV :-) JD : Ooh that’s not good Must have missed that in the template I copied. I’ll correct that now Baptiste : are beam formed data 5nançay) and single dish data the same ? AL : we will think about it. MK : Most beam forming is used to produce dynamic spectra in some form, isn’t it? MK : Interferometers typically produce cubes.