Axes of Discovery: the Time Variable Universe
Jim Cordes, Cornell University
Heraclitus “You don’t observe the same universe twice”
Discovery frontier for  + CR + GWs
– Less so at high energies
• BATSE, RXTE/ASM, Beppo/Sax, SWIFT, GLAST, etc.
– More so for optical, radio
GLAST: full-sky survey every three hours for 1+ years
Optical
• PanSTARRS = Panoramic Survey Telescope and Rapid Response System
• LSST (LST) = Large Synoptic Survey Telescope (rename to OSST!)
• Science goals: includes transients on minute  10 yr time scales
Radio
• Phase space:
– Knowns: already a very rich set
– Hypotheticals
Radio Synoptic Survey Telescopes
• Survey metric for transients
• The mid-frequency SKA as the RSST
2 Oct 2007
Jim Cordes
Modern Radio Universe
1
Why Radio Transients?
No comprehensive survey of large phase space
• Need large AT (area, solid angle and time coverage)
• ns to years
• complex structures in the frequency-time plane  HPC processing
Nature can produce cheap radio photons via coherent radiation
processes (N2 vs N) so detectable to great distances
Fast transients are linked to extreme matter states (t < 1s)
… or ETI
Counterparts to known source classes
• Prompt radio bursts from GRBs
• Gamma-ray quiet, radio-loud GRBs
Expect new source classes (ETI, evaporating BHs, particle
events)
Beacons for probing the cosmic web and fundamental constants
• Intervening plasmas (IPM, ISM, IGM)  dispersion, scattering, scintillation
• Photon mass, charge from measured dispersion law (de Broglie 1940)
2 Oct 2007
Jim Cordes
Modern Radio Universe
2
Galactic Center Transients
GCRT J1745-3009
Hyman et al. 2005
2 Oct 2007
Bower et al. 2005
Spk ~ 80 mJy
Jim Cordes
SD2 ~ 5.1 Jy kpc2 W ~ 100 d
Modern Radio Universe
3
PMB Single Pulse Search
McLaughlin et al. (2006)
 Roughly 1/3 of all psrs
detected with FFT also
detected in SP search
 Some objects detected only
in SP search
J1840–0809
 Wide range of single-pulse J1840–0815
properties apparent
P = 0.96 s
P = 1.1 s
-3
DM = 353 pc cm-3
 Galactic pulsar population DM = 225 pc cm
may be much larger than
RRATs: rotating radio transients
previously thought
(Andrew Lyne’s talk)
2 Oct 2007
Jim Cordes
Modern Radio Universe
4
Pulsar Survey with Arecibo Multibeam System (ALFA)
(Arecibo ~ 10% SKA)
Detection of a strong
pulsar amid RFI
2 Oct 2007
Jim Cordes
Detection of a weak millisecond
pulsar in beam 1
Modern Radio Universe
5
2 Oct 2007
Jim Cordes
Modern Radio Universe
6
astro-ph/0605145
• Asteroid disk from SN
fallback material
• Only ~ 10-4 Earth
masses needed to provide
enough material to perturb
coherent radiation from a
pulsar over its 10 Myr
lifetime
• Asteroids evaporate at
~109 cm from the NS and
trigger or quench pair
production from gaps in
the magnetosphere
• Expect induced torque
fluctuations
2 Oct 2007
Jim Cordes
Modern Radio Universe
7
Brown dwarf pulsations
Hallinan et al. 2007
Time series of the radio emission
detected with the VLA from the
M9 dwarf TVLM 513-46546.
Every 1.958 hours a periodic pulse
is detected when extremely bright,
beams of radiation originating at
the poles sweep Earth when the
dwarf rotates.
2 Oct 2007
Jim Cordes
Artist's impression of a brown dwarf
with "super-aurorae" at its magnetic
poles, causing the pulsed radio
emission. (Credit: Copyright National
University of Ireland, Armagh
Observatory,
National
Radio
Astronomy
Observatory,
United
States Naval Observatory & Vatican
Observatory, Arizona)
Modern Radio Universe
8
Bursting radio emission from magnetars
Correlated torque and radio variations
Flat spectrum (not pulsar like)
Camilo et al. 2007
1E 1547-5408
2 Oct 2007
Jim Cordes
Modern Radio Universe
9
Crab giant pulses: the highest brightness temperatures known
“nano shots”
2 Oct 2007
Jim Cordes
Modern Radio Universe
10
Frequency-time structure in a Crab Giant Pulse
Eilek & Hankins 2006, Hankins & Eilek 2007
Arecibo data
2 Oct 2007
Jim Cordes
Modern Radio Universe
11
Phase Space for
Transients:
SpkD2 vs. W
W = pulse width
or characteristic
time scale
2 Oct 2007
Jim Cordes
Modern Radio Universe
12
Filling phase
space with
hypothetical
new discoveries:
Prompt Gamma-ray
emission
Evaporating black
holes
Maximal giant pulse
emission from
pulsars
ETI’s asteriod radar
What else?
2 Oct 2007
Jim Cordes
Modern Radio Universe
13
Detection limits
for the SKA:
SpkD2 >threshold
 Prompt GRBs and
GRB afterglows easily
seen to cosmological
distances
Giant pulses detectable
to Virgo cluster
Radio magnetars
detectable to Virgo
ET radar across Galaxy
2 Oct 2007
Jim Cordes
Modern Radio Universe
14
Planetary Radio Emission
Cyclotron maser emission
Radiometric Bodes Law
Desch & Kaiser 1984
Lazio et al. 2003
Bastian, Dulk & Leblanc (2000)
2 Oct 2007
Jim Cordes
Modern Radio Universe
15
Science Express
27 Sep 2007
W ~ 4.6 ms ( / 1.4 GHz)-4.80.4
DM ~ 375 pc cm-3
Steep spectrum (=-4)
2 Oct 2007
Jim Cordes
Modern Radio Universe
16
Detection limits
for the SKA:
SpkD2 >threshold
 Prompt GRBs and
GRB afterglows easily
seen to cosmological
distances
Parkes SP
(if D>500 Mpc)
Giant pulses detectable
to Virgo cluster
Radio magnetars
detectable to Virgo
ET radar across Galaxy
2 Oct 2007
Jim Cordes
Modern Radio Universe
17
Prompt Radio Bursts from -ray Bursts
Scale from -ray fluence:
Implied flux density:
Reasonable?
Compare with maximal pulsar energy losses
n+1 ~ 3.5
Can radiation get out? (Macquart 2007)
Induced Compton and Raman scattering
Long bursts: hypernovae
 dense plasma from pre-SN stellar wind, immersed in SF region
 perhaps no emergent radio emission
mitigating effect: radio coherent, upboosted photons incoherent
Short bursts: merging NS-NS, NS-BH
 vacua that high-brightness radiation can get through
 plausible radio bursts through reactivation of magnetosphere
2 Oct 2007
Jim Cordes
Modern Radio Universe
19
Isolated pulsar
Re-activation of
pulsar action in
mergers?
Hansen & Lyutikov 2000
Lyutikov 2006
2 Oct 2007
Jim Cordes
Modern Radio Universe
20
Transient Surveys
• Fast
• Too fast to be sampled by raster scanning (<1 day)
• Sub-second transients:
– Produced by compact sources (size < ct)
– Coherent radiation
– Influenced by diffractive interstellar scintillations
 imposed -t structure on intrinsic signal
• Sampled by “staring” for long dwell times
• Large solid angle coverage needed for rare events
– Likelihood of detection
– Completeness level
• Very high data rates for full FoV analysis
• Slow
• Durations long enough to be handled with imaging raster
scans
2 Oct 2007
Jim Cordes
Modern Radio Universe
21
Survey Metrics
“Survey speed” for Steady Sources:
• Payoff = number of objects detected to some flux
density level in some fixed amount of time
• Steady sources, homogeneously distributed, etc.
» Integration time per source = dwell time in survey 
• Get same metric by looking at rate at which volume or
solid angle surveyed:
 SS = FoV (A/T)2
• Processed bandwidth enters in linearly
 SS = FoV (A/T)2 B
• Extended survey metric that includes other factors:
fc = fraction usable antennas
NFoV = number of pixels (PAFs)
Nsa = number of subarrays
m = signifcance level (min. S/N)
2 Oct 2007
Jim Cordes
Modern Radio Universe
22
Survey Metrics
Survey speed for Transient Sources:
W = transient duration,  = dwell time in survey scan
Slow transients: integration time = 
Fast transients: (days  ns) integration time = W
 New factor on survey metric that accounts for
integration time and time-capture probability
FoMTS = FoMSS  integration time factor  Pt
•
•
•
•
K(a,x)  1
Integration
time factor
x=/W
Probability that  1 event occurs
from source when pointed at
a = W,  = event rate/source
See SKA memo by JMC to appear at www.skatelescope.org
2 Oct 2007
Jim Cordes
Modern Radio Universe
23
$$
$$$$$
Sensitivity vs.
instantaneous
FoV
For:
1 GHz
1 sec integration
0.3 GHz bw
¢¢¢¢¢
2 Oct 2007
Jim Cordes
Modern Radio Universe
25 K, 60%
24
Wide FoV Non-Imaging Surveys
• Pixelization of the field of view
• Correlation approach favored over beam forming
• Number of pixels
• Sky coverage
• Raster scanning (slow transients)
• Staring (fast transients)
• Analysis
• Full search analysis on each pixel
• Extensive for pulsars and fast transients
– E.g. 1024 frequency channels  64 s samples from each pixel
• Frequency-time plane analysis for all cases to discriminate
RFI from celestial signals
2 Oct 2007
Jim Cordes
Modern Radio Universe
25
Example Synoptic Cycle for SKA
A 10-day total cycle: variable scanning rates
– Fast for extragalactic sky (away from Galactic plane)
E.g.
• 1 deg2 single pixel FoV
• Full sky survey (80% of 40,000 deg2)
• Tscan = 5 days
• T ~ 10 sec = time per sky position
• Smin ~ 15 Jy at 10 with full sensitivity and on axis
• Multiple pixel systems (PAFs) increase sensitivity (for fixed total time)
• Subarrays reduce sensitivity but speed up the survey
–
–
–
–
–
Slow for deep extragalactic fields and Galactic plane
Galactic center: staring mode
Repeat scans many times
Break out of scanning mode for targeted observations (10%?)
Break out for targets of opportunity
Issues for pulsars (~steady amplitudes):
– Need minimum contiguous dwell time for Fourier transforms (e.g. 100 – 1000 s
for large-area blind surveys)
– Need frequent re-observation coverage for long-term timing followup
Calibration requirements
Are there solutions for HI, pulsars, transients, SETI and magnetism?
– Yes (to zeroth order)
2 Oct 2007
Jim Cordes
Modern Radio Universe
26
Summary
• A complete inventory requires attention to the time domain on many
scales
• A multi-wavelength enterprise but some radio-unique areas too
• Target classes for transients
Causes:
• Relativistic objects (stellar, AGN)
• internal instabilities
• Planets, brown dwarfs
• ETI
• Evaporating black holes? What else?
• Keys to discovery at radio wavelengths
• Time-frequency (t-) analysis
• Wide-field sampling ()
• Low-mid-high sensitivity (smin)
• extrinsic triggers
• external modulations
• lensing
• scintillations
• Blind surveys of the radio sky
• Expand AeT
• Comprehensive matched filtering (computing)
• Start now
Arecibo, GBT, Parkes,WSRT,EVLA ATA,LOFAR, LWA,MWA,ASKAP,MeerKAT
SKA
• Fast transients: obtain & process data as in pulsar surveys, SETI +?
• Slow transients: raster scan the sky repetitively
• Develop the SKA as a Radio Synoptic Survey Telescope (RSST)
2 Oct 2007
Jim Cordes
Modern Radio Universe
27
Extra Slides
2 Oct 2007
Jim Cordes
Modern Radio Universe
28
Arecibo/ALFA
(30% of sky in
2000 hr)
high rate
SKA/PAF (80%
of sky in 5 days)
or
SKA/SP (80% of
sky in 5 days)
“W” limited
low rate
2 Oct 2007
Jim Cordes
Modern Radio Universe
29
Science and Implementation of RSST
• SKA mid-frequencies: 0.3 to >3 GHz
• Hydrogen
0.3 – 1.4 GHz
» (galaxy evolution, dark energy)
• Pulsars and Gravity
» (GR, GWaves, EoS, Magnetospheres)
• Transients (ns to years)
~0.8 – 3 GHz
0.3 – 8 GHz
» Relativistic objects
» Exoplanets
» SETI
• Magnetic Universe
1 – 2 GHz
• Synoptic and Commensal Surveys
•
•
•
•
Sensitivity + Wide FoV + Frequency-time flexibility
Spectral line + pseudo-continuum + time-domain surveys
Need multiple backend processors
Compatibility issues (configuration, cadences, scan types)
» HI galaxy evolution
» Pulsar/transient full-FoV search
2 Oct 2007
Jim Cordes
~10 km baselines
< 1 km
Modern Radio Universe
30
Small and Big Science
• Discovery space can be explored with both low
sensitivity and high sensitivity instruments
•
•
•
•
W space: ns  years
 space: MHz  -rays
rate space: >>1 s-1 to 1 d-1 hemisphere-1
flux densities: Jy to GJy
• Fast transients require matched filtering in the -t plane
• High-resolution, high-performance processing
•
•
•
•
Minimal systems can make useful observations
Low-f: LOFAR, LWA, MWA
Mid-f: Arecibo, GBT, EVLA, ATA, ASKAP, MeerKAT
High-f: GBT, EVLA, ALMA?
2 Oct 2007
Jim Cordes
Modern Radio Universe
31