Transcript Save the Sky: Adventures in Sky Monitoring

Save the Sky:
Adventures in Sky Monitoring
Robert J. Nemiroff
Who am I?
Most cited science papers:
GRBs: time dilation, cosmology, lens searches
Microlensing: finite source size effects, AGN BLR probe
Favorite science papers:
On the Probability of Detection of a Single Gravitational Lens
(1989)
Visual Distortions Near a Black Hole and Neutron Star (1993)
Toward a Continuous Record of the Sky (1999)
Tile or Stare? Cadence and Sky-monitoring Observing
Strategies That Maximize the Number of Discovered
Transients (2003)
Who am I?
(Know your IAS visitors)
Web:
Black hole movies at:
http://antwrp.gsfc.nasa.gov/htmltest/rjn_bht.html
GR correct! (Could make another IAS talk)
Astronomy Picture of the Day at:
http://apod.nasa.gov/
NASA’s top-ranked site!
Who am I?
Cool ideas I’d like to better explore:
(Want to collaborate?)
Placing a satellite at 50 AU to use the Sun as a
transparent lens
Placing limits on magnitude of lower order w
components
Estimating if Swift/GLAST will help find GRB
lensing
Save the Sky
What happened in the sky last night?
Supernova? Nova? Eta Carina flare?
GRB afterglow? Undocumented flash?
Flurry of sporadic meteors?
Clouds obscure your remote observing?
Cirrus affect data on Jan 22 at KPNO?
Are clouds rolling in just now?
Is last night’s sky gone forever?
Save the Sky
Popular Name:
The Night Sky Live Project
Web address: http://concam.net
Deploys CONtinuous CAMeras
(CONCAMs)
CONCAM: Objectives
Primary Science
Unprecedented temporal monitoring for GRB OTs, meteors,
variable stars, comets, novae, supernovae
Support Science
Unprecedented ability to act as instantaneous cloud
monitors, archival cloud monitors, generate all-sky
transparency maps, all-sky emissivity maps
Education / Outreach
Unprecedented ability to show your class last night’s (real)
sky, archival skies, monitor meteor showers in real time,
show educational sky movies, run educational modules
CONCAM Locations
Save the Sky: 4 CONCAM locations
Kitt Peak
Mauna Kea
Mt. Wilson
Wise Obs.
CONCAM: Hardware
CONCAMs are essentially fisheye lenses
attached to CCDs run by a PC computer and
connected to the internet. CONCAMs do not
move - they are completely passive.
Most simply put: light comes in the top, electricity comes in
the bottom, and data flow out the bottom.
In building CONCAMs, we have three
montras:
“If it moves, it breaks.”
“The lens IS the dome.”
“Don’t spend 90% of your time trying to get 10% more
images.”
CONCAM: Data
All recent images are available through
http://concam.net
All data are free and public domain.
All FITS and JPG data are archived to DVDs
(previously CDs).
Each CONCAM node generates about 500Mb
of raw image data per night.
Higher level data products (e.g. photometry)
are now being generated in real time for
some CONCAMs.
CONCAM Scientific Milestones
First CCD device to image the position of a gamma-ray burst during the
time of the gamma-ray burst trigger (#1: GRB 001005)
Most complete and uniform coverage of a meteor storm: the 2001
Leonids
Most complete light curves for hundreds of bright variable stars starting
from May 2000, when the first CONCAM was deployed on Kitt Peak.
First devices to give real-time optical ground truth for the whole sky in
support of major astronomical telescopes, including Gemini North,
Keck, Subaru, IRTF, SpaceWatch, Wise, ING 4-m, Mayall 4-M, SARA,
and WIYN.
In May 2003, fisheye night sky webcams now image most of the night
sky, most of the time. For example, were SN 1987A to go off
tomorrow, there would be a good chance that a CONCAM saw it.
Tile or Stare?
A sky monitor’s classic conundrum
Sky monitoring increasing
Current Projects (see BP webpage: abridged, expanded)
CONCAM
KAIT
LINEAR
LONEOS
LOTIS
MEGA
NEAT
RAPTOR
ROTSE
Spacewatch
STARE
SuperMACHO
TAOS
YSTAR
R. J. Nemiroff
A. Filippenko
LINEAR team
T. Bowell
H. S. Park
A. Crotts
E. Helin
W. T. Vestrand
C. Ackerloff
R. S. McMillan
T. M. Brown
C. Stubbs
C. Alcock
Y. I. Byun
Tile or Stare?
Likely future sky monitoring projects
include (much abridged):
Pan-STARRS
LSST
GLAST
N. Kaiser
A. Tyson
P. F. Michelson
Tile or Stare?: Assumptions
Generic case considered here:
Transients are discovered and confirmed on a time-contiguous
series of exposures
Sky is isotropic
Effective apparent brightness distribution of transients N(l) is
already known
Once discovered, transients are handed off to a separate followup telescope
“Tile or Stare” & tiling cadence determination important for:
microlensing, GRB OTs, supernovae, planet detection, binary
star eclipses, stellar flares, blazar flares, QSO flares, Near Earth
Objects, comets, meteors & more ...
Tile or Stare?
The Two Key Power Indices:
, 
Variables:
N: effective apparent cumulative brightness
distribution of transients
ldim: apparent luminosity at obs. limit
te: exposure time
At the observation limit, quantify:
N  ldim (low background:   -1)
ldim  te (high background :   -1/2)
N  te
Tile or Stare?: A Mathematical Optimization
Find N(l) from existing observations (l: apparent brightness)
Find l(te) from detector, noise, and backgrounds (te: exposure
time)
Compute N(te) -- might be conveniently parameterized in terms
of power-law indices  & 
Estimate total time of campaign: tc (exact value usually not
important)
Find grand total expected transients during campaign: Ng
Write Ng is terms of treturn, the time it takes for a survey to
return to a given field (i.e. cadence). Read, down and slew
times enter here.
Compute dNg/dtreturn, find solutions to dNg/dtreturn=0.
Find treturn that best maximizes Ng.
Save the Sky: Cadence
Tile or Stare: Cadence
Tile or Stare: Cadence
Tile or Stare? Decision Summary
If, during exposure, the rate that transients come over the
limiting magnitude horizon is increasing fast enough (  >
1), then stare should be preferred.
If, on the other hand, the rate that transients come over the
limiting magnitude horizon is not increasing fast enough ( 
< 1), then tile should be preferred.
Usually the best tiling cadence is the duration of the transient,
since a faster tiling cadence will waste effort on transients that
have been previously discovered, while a slower tiling cadence
will miss transients occurring in other fields.
If, however, the duration of the transient is comparable to the
cumulative read-out and/or slew times during a sky-tiling, then
a mathematical maximization as described in the preprint will
find the most productive cadence.
Tile or Stare? SuperMACHO
Objective: maximize microlensing transients discovered
LMC N(l) has  < 1: tile beats stare for identical fields
what cadence?
LMC not isotropic: fields with highest N(ldim) preferred
N(l) may change with seeing or be better determined with time
Therefore, choosing the next field to observe is very complicated -not unlike a chess game. Optimization might involve real-time
Monte-Carlo simulations.
Field return rate still attracted toward transient “duration of
interest”
faster cadence inefficiently re-discovers known microlenses
(competes with field richness at ldim)
“duration of interest” may be the microlens rise time: ~ two weeks,
although microlens rise times have wide variety of durations
Tile or Stare?: LSST
Objective (example): maximizing Type IA supernovae discovered
Sky essentially isotropic (out of Galactic plane)
N(l):  > 1 for I < 24: stare preferred
effectively creates a minimum observation time per field
N(l):  < 1 for I < 24: tile preferred
what cadence?
Return time (cadence) optimized at the “duration of interest”
faster cadence inefficiently re-discovers known supernovae
slower cadence inefficiently misses supernovae in neglected fields
“duration of interest” could be rise time of SNe: ~ 15 days (1+z)
Different cadences will optimize discovery rates for different
transients
might have Guest Investigators (GIs) program where GIs change
filters and cadence to optimize discovery rate of GI-preferred
transients
Tile or Stare?: GLAST
Objective: maximize blazars (quiescent phase) discovered
GLAST’s survey mode constrains it to point away from the
Earth, but rock at some cadence between the N&S Celestial
Poles.
N(l) away from Galactic Plane:  > 1: stare
stare = GLAST Deep Field (GDF); should maximize detections
stare only possible at NCP, SCP or during pointing mode
GDF exposures should end if/when faint blazars saturate
( drops below unity)
N(1) in Galactic Plane:  < 1: tile
GDF strategy inefficient in Galactic Plane
quiescent nature allows co-adding at any time, cadence
unimportant
, , GDF existence, GDF location are energy dependant.
Support