Richard Armstrong

Commensal Transient Searches with KAT-7 and MeerKAT

1. Richard Armstrong – Cape Town. Commensal Transient Searches with KAT-7 and MeerKAT
2. Overview   ThunderKAT: why observe commensally?   Design goals for a MeerKAT commensal transient pipeline   System Details   Project schedule
3. ThunderKAT: Extreme Physical Environments   Synchrotron radio transients:   Relativistic jets (microquasars), jets in WDs, outflow from novae   Extragalactic X-ray binaries   Supernovae   Gamma-Ray Bursts,   Magnetar outbursts   ThunderKAT Strategy   Transient detection and monitoring (commensal and pointed)   Follow up of transients over broad frequency range with other instruments   ThunderKAT transients to be reported to the multi-wavelength community via a standard protocol: VOEventNet (SkyAlert)
4. Commensal Observation   Observe simultaneously with ALL other large science surveys.   Second goal of the ThunderKAT project: to search both commensally and in a targeted way for new radio transients, and to classify / identify and report on these as rapidly as possible   Why? Extend the effective number of hours allocated to ThunderKAT to all allocated hours:
5. Commensal Observation   MeerKAT TAC recommendation: pursue commensal observing with all other projects.   Commensal observing designed in to the Large Surveys (70% of MeerKAT time) from the beginning.   Discovery machine:   public arcsec localisation to thousands of new events across the whole southern sky   prime source for first discovering extreme objects for both MeerKAT and global multi- wavelength follow-up .
6. Commensal Example   Example: commensal observing with MHONGOOSE and Fornax MHONGOOSE: Nearby galaxies with distance in the range of 4 - 10 Mpc   200 h per galaxy, e.g. 25 epochs: 8 hours/epoch (for 30 galaxies)   Fornax cluster: Cluster environment at a distance of 17 Mpc
7. Commensal Observation   Detect transience during other surveys   either astrophysical or   instrumental   May be able to distinguish between these.   See talk by Ian Heywood   Independent imaging pipeline
8. The Tough Bits   Real-time, autonomous imaging, detection, flux correction and spectral classification (control your laughter).   We do not have the choice of fields and cadences.   Real time? Not inherently a hard requirement. But otherwise when else shall data be processed?
9. The Lucky Bits   We (ThunderKAT) don’t have to store raw data. MeerKAT and KAT-7 do this for us.   Transient detection framework built over many years by our friends at LOFAR TKP   But: may need a lot of modification to work correctly for KAT-7; still evolving itself.   May eventually need to start from scratch again (with lessons learned)   KAT is an open, helpful facility to work with.
10. Design Strategy   First Goal:   end-to-end pipeline using existing components to reduce time to measurements.   This is why the LOFAR Transients Pipeline (TRaP) was chosen:   many years of development, benefit from lessons learned by others.   Risk:   it is designed for a telescope with different parameters.   Data rates of LOFAR greater than KAT-7 and MeerKAT
11. Design Strategy   MeerKAT Integration:   Interfaces remain the same (SPEAD, KAT-live)   Interfaces extend to future KAT iterations (MeerKAT-16, MeerKAT-64), thus minimal re-work
12. System Diagram
13. Some Details KAT Correlator (and future versions) SPEAD Source Catalogue Database Imaging SPEAD Source Finding Calibratio n/Flux Measure ment Control Computer ML (machine learning) System: decide if Important VOEvent BBRA (black belt radio astronomer) Modifiable/ Replaceable Software HDF5 data storage Calibration & Imaging Convert MS, FITS, etc Global multi-wavelength Follow-up Real-time Operation ThunderKAT Server KAT live Imaging, Source Detection and Classification
14. Project Schedule   KAT-7   May: arrival of ThunderKAT server   July: tests of LOFAR TraP with existing KAT-7 data set.   July/August: acquire storage (?)   End of August: first tests with live data from KAT-7   October/November: integration with VOEvent Network   MeerKAT   Move KAT Transient Pipeline closer to the instrument to reduce data-rate over the Cape-Town  Carnarvon link.   Further details as MeerKAT schedule evolves
15. Early pipeline results (CASA-based)
16. Image plane
17. Early pipeline results   A far cry from automatic imaging   Targeted observation; no source detection   Does expose the issues with autonomous pipeline   How to detect when the reduction has not converged?   Extracting measurement noise from the image is non-trivial.   Does a visibility-plane transient system make more sense? Why is it important to image? New algorithms required?
18. Summary   Commensal observation increases the effective time allocation of ThunderKAT   We’ll Initially use the LOFAR TraP (Transients Pipeline)   Schedule for commensal observation with VOEvent notification by year-end
19. End