SWAN : Sky Watch Array Network ---- A Strategic Initiative

1. SWAN : Sky Watch Array Network---- A Strategic Initiative --desh Raman Research Institute, Bangalore, India Pre-ASI workshop on Transients @NCRA, Pune. 16 Feb 2015

2. 0. The radio “transient story” so far1. Transient searches : ongoing effort, and possible optimizations 2. SWAN: What, Why and How3. Summary=====This story starts with.. the so-called “Lorimer Burst” (2007) reported from the Parkes Telescope.... detection of a dispersed pulse

3. Frequency Intensity Intensity Effects of Dispersion and its Correction ٧ = 1420 MHz, ∆٧ = 32 MHz ∆t/DM << 1 msec ; ٧ = 35 MHz, ∆٧ = 1 MHz, ∆t/DM ~ 200 msec ; • Incoherent (Post-detection) dedispersion: Signal time sequence from each frequency channel is shifted backwards by a dispersive delay ∆t 3

4. “Radio Bursts with Extragalactic Spectral Characteristics Show Terrestrial Origins” .....Sarah Burke-Spolaor et al 2010

5. Thornton et al. 2013 If real, expect several hundreds per day !! ..~20 till now Lorimer burst like recent Fast-Radio-Bursts (FRBs)

6. Extra-galactic origin ?!!

7. Now, detection of similar bursts also from Arecibo.

8. Potential Sources of Fast Radio Transients • Pulsars : ... Found so far > 2000 Expected number is much greater than this (The Crab Pulsar was found due to its giant pulses) 2. RRATs : A new category of radio sources ( McLaughlin, M. A., 2006) Not explored much at low frequencies 8

9. Potential Sources of Fast Radio Transients 1. Pulsars Giant pulses (coherent emission from sometimes a beach-ball size region <-- nanosecond) 2. RRATS 3. Radio flares from Brown dwarfs Jupiter Solar bursts Active stars AGN outbursts (Extra-galactic, higher frequencies) 4. Many other known and unknown ??! sources Recent reports on Fast Radio Bursts (FRBs) ... argued to be of extragalactic origin (large DMs) ! 9

10. On-going efforts • Parkes-multibeam • Arecibo-ALFA (multibeam) • MWA • LOFAR • LWA • VLBA • GMRT • ….. etc.

11. 1. General considerations for optimum detection: a) targeted search b) blind search might differ in size of the instantaneous field of view. Common to both is Matched Filtering... parameter space:dispersion measure, pulse width/shape (and period) Here, we focus on non-recurring fast transients, which are hard to catch, and even harder to confirm.

12. Fast transients: signal necessarily from compact sources, broad-band (also coherent across BW = 1/temporal_width), and most likely highly polarizedNOTE: sensitivity increases LINEARLY with coherence bandwidth, which could be high here, unlike for steady radio noise

13. a) targeted search: detection probability dp <-- sensitivity <-- collecting area (A), bandwidth, pulse-width as long as beam-size >= source-size, higher A and longer look --> higher dp

14. b) all-sky (blind) search: sensitivity <-- A; probed-distance <-- sqrt(A) sky coverage at any time <-- N_beams/Aprobed volume <-- N_beams/sqrt(A)The benefit of large A can be realized only with the correspondingly larger number of beams.(Of course, relevant luminosity distribution of the targets of interest is also to be included).

15. These considerations assumed filled-apertures. Dilute apertures have poorervolume coverage for fast transients. Use of multiple apertures to cover different directions appears more rewarding, compared to combining themas an incoherent array.... ....as far as “sampled volume” is concerned. Bandwidths: for delta_f < BW_coh < BW, sensitivity <-- sqrt(BW*BW_coh)

16. Radio Frequency Interference (RFI) Detection & Mitigation First order detection/mitigation happens naturally i.e. Broadband RFI show up at 0.0 DM-value Identify deviations from the otherwise smooth spectrum as narrowband RFI .. and exclude such spectral channels Monitor ratio of the expected to the observed standard deviation of intensity variation in each of the channels; RFI will show a dip 16

17. Radio Frequency Interference (RFI) Detection & Mitigation Can antennas filter pre-specified RFI bands ? Yes. Can polarization filter be used, i.e. Can Removal of polarized component help ? No. Risky !... since your transient signal Is likely to be polarized. 17

18. How to ensure immunity to RFI ? Go to remote places free of RFI ? e.g. MWA. ... but no escape from satellite signals .... Moon reflects man-made RFI Use of multiple locations.

19. How to ensure immunity to RFI ? Appeal to large span in frequency, or rather span in wavelength-square.. to exploit the dispersion characteristics of the sky signal. Even the expected frequency dependence can be checked for, i.e. f^-2, as a criterion.

20. Yogesh Maan + from Gauribidanur Telescope

21. Large Bandwidth v/s wide spectral span • Detection probability increase is slow • (BW)^(1/4) Same resource as required to handle wider BW can be utilized better in other ways. However, wider span offers distinct advantage in more than one ways. Hence, multi-band operation is appealing !

22. A sample from observations withRRI-GBT multi-band receiver @Green Bank

23. Single-pulse detection • Single pulse search technique: Methodology: 1. Dedispersion at a range of DMs 2. Matched-filtering for different pulse widths 3. Thresholding for significance assessment 4. Diagnostics • Radio Frequency Interference : Detection & mitigation 23

24. Frequency Intensity Intensity Effects of Dispersion and its Correction ٧ = 1420 MHz, ∆٧ = 32 MHz ∆t/DM << 1 msec ; ٧ = 35 MHz, ∆٧ = 1 MHz, ∆t/DM ~ 200 msec ; • Incoherent (Post-detection) dedispersion: Signal time sequence from each frequency channel is shifted backwards by a dispersive delay ∆t 24

25. Signal to Noise (S/N) Dispersion Measure Spacing of DM-values Δ DM  ƒ3/Δƒ  Δt Example detection by dedispersing over a trial DM-value range Data taken from observations in the direction of pulsar B1133+16 25

26. Matched Filtering Smoothing time = width of the pulse=> Maximum Signal to noise Ratio Most Dominant Factor which decides the Pulse-Width at 35 MHz : Interstellar Scattering • Convolution of the intrinsic pulse with a one-sided exponential function • Scattering time-scale ~ ƒ-4 ×DM2 26

27. B0834+06 27

28. 2. Sensitivity Periodicity search needs a number of pulses present in the data to be searched SPS is sensitive to even a single bright pulse, even if buried in RFI. More than one transient sources present in the beam can be detected simultaneously (using MST-Radar antenna) 28

29. Strategic Initiatives: Indian SWAN (Sky Watch Array Network) 1) To facilitate and conduct searches and studies of fast (typically of sub-second duration) and slow transient radio radiation originating from astronomical sources. 2) To facilitate and conduct high angular resolution imaging of discrete galactic and extragalactic sources at low radio frequencies. (down to a few tens of milli-arc-seconds) 3) To train, involve and provide hands-on experience to a large number of undergraduate and postgraduate students in all aspects of the SWAN, through their direct and active participation starting from the design stage to research using the array network.

30. Why ? Growing realization about the transient sky, opening up an entirely new dimension of astronomical exploration with a huge potential for yet unanticipated discoveries. Similar initiatives at optical wavelengths have yielded exciting results, but corresponding view at radio range is lacked. There have been recent reports of detection of fast radio bursts, believed to be of extragalactic origin, but with little possibility of confirming them or studying their origin. The reported detection imply a high rate of such events, but no optimized setup exists yet anywhere in the world to search a large volume of the space with required sensitivity to detect such signals routinely and to enable a proper study. Indian-SWAN will fill this gap by providing capabilities that are carefully optimized together for extensive search, reliable detection and localization of energetic radio transients (including the ones that do not recur), and for their possible follow-up.

31. Why ? While cm and mm wave studies of radio sources have been possible at angular resolutions of even sub-arc-second level, meter-wave studies have not yet been possible at any comparable resolution. There is a need thus for such a very long baseline array, particularly in India, complementing the capabilities of the GMRT at the low radio frequencies. The proposed Indian-SWAN will extend our capabilities for studies at these frequencies in both sensitivity and angular resolution, by significant factors.

32. Why ? An overwhelming majority of the bright and motivated students still remain unexposed to the exciting developments and research opportunities in radio astronomy. Due to the lack of exposure at an early stage, much of the talent is missed to be attracted to graduate studies and research in astronomy. To provide extensive opportunities to a large number of undergraduate and post- graduate students at the Indian technology and science institutes, as well as at some universities to be directly involved in the realization of the SWAN, including in the design, and also in the operation/usage of the setup for astronomical observations, as well as the follow-up/research. Thus providing them a state-of-the-art facility in their back-yards, also for trying out their own ideas. These opportunities have the potential to seed significant growth of future generations of Indian radio astronomers pursuing active research, including with the future large telescopes (e.g. SKAs).

33. How ? Co-ordinated network of arrays: ~ 1000 sq. m array area at each location & a decade frequency range: 50-500 MHz. Phase-0 : a narrow-band setup for proof of concept, with 8 independent stations, using most of the available hardware (e.g. GBT receiver, MWA tiles, etc. ). First at Gauribidanur, then relocation at 8 different sites. Will be used for serious observation, including VLBI with GMRT. Phase-1: increase the aperture size and spectral coverage for the 8-station setup, providing most of the observational modes. Phase-2: the network will be swiftly expanded to the full set to reach the aimed watch (or survey) sensitivity and volume coverage, etc. A collective effort to develop the SWAN with as many of the 40+ technology and science institutes, as well as universities across India. Appropriate schools/workshops to prepare for student involvement.

34. MWA VCS Tile (4 x 4 dipole array) Beamformer Receiver (for 8 Tiles) MWA Correlator (128 Tiles, 3072 channels)

35. @ Green Bank

36. Rotman Lens as a beamformer Array Ports beam Ports

37. Summary • In Blind search : collecting area advantage is retained through use of multiple beams • Benefits at Low-frequencies and/or with large spectral span • Multiple stations crucial • SWAN • Exciting times ahead ! • Thank you for listening.

39. For my Indian colleagues: • Happy to pursue an initiative... involving undergrad students and teachers at the several Science and Technology institutes across India.... to create a multi-station setup optimized for radio transient search and follow-up. (TBD, but might have a decade spectral span, and many tens of stations, etc.) • If you wish to be a part of this venture, contributing a fraction of your time & effort (say, > 6 months, spread) over the next 5 years, then please let me know via e-mail desh@rri.res.in • Thanks.

40. In contrast, for survey of non-transient sky (or non-varying over the total look time T)If T0 is integration time per field,sensitivity <-- A.sqrt(T0); sky coverage in T <-- (N_beams/A).(T/T0)volume/T <-- (N_beams/sqrt(A)).T0**-(3/4)(for fixed sensitivity: volume/T <-- N_beams.A)for fixed A: volume/T <-- N_beams/ T0**3/4

41. Arecibo related development Arecibo: GALFACTS L-band, seven pixels, 300-MHz BW, spectral-resolution -> 1 MHz time resolution 1 msec. Full Stokes Meridian nodding mode, basket-weaving