Status of the IGEC2 joint burst search

1
Status of the IGEC-2 joint burst search
G.A.Prodi, INFN and Univ. of Trento, Italy on
behalf of the IGEC-2 Collaboration
2
outline

• IGEC-2 agreement and burst search
• Comparison with IGEC-1 burst search
• Methodology of network analysis
• detector performances
• Features of the first exchanged data set
• VERY preliminary results of network analysis and
  hints on the IGEC-2 expected performances

3
IGEC-2 agreement Aspen Conf., Jan. 17th,
2005document available at http//www.auriga.lnl.i
nfn.it

• Agreement for a joint search for gravitational
  waves based on our previous experience (IGEC
  1997-2000)
• Participating groups ALLEGRO, AURIGA and ROG.
• Leadership AL, AU and ROG spokespersons, who
  unanimously decide actions/agreements. Stronger
  coordination of joint observations.
• Open to agreements for joint searches with other
  groups running detectors
• First search time coincidence search for bursts
• 2004 May ? 2005 Feb
• Task force, coordinator G.A.Prodi,
• vice-coordinators W.Johnson and M.Visco.

4
Overview of data analysis methodology

• New blind data exchange for a blind data
  analysis
• Rigid time shifts has been secretly added by each
  group and will be circulated only when the
  analysis procedure is agreed in detail
• Network analysis based on IGEC-1 experience use
  a priori information to improve the network
  search (signal template, testing source
  locations, common search thresholds on
  amplitudes, etc.)
• Nfold-time coincidence search with adapting order
  N
• a priori control of false dismissal (conservative
  bound).
• Data selection, time coincidence search and
  accidental coincidence estimation in the footpath
  of IGEC-1

5
overview of data analysis

• New data analysis challenges
• cross validation of exchanged data
• commitment to perform coordinated injections of
  software signals (Mock Data Challanges).
• more information exchange on the performances of
  the detectors and data analysis procedures
• systematic use of directional searches (maps of
  the sky)
• multiple trials
• confidence belt contruction and unfolding
• control of the False Discovery Rate
• development of more powerful code for the network
  data analysis
• extension to non parallel detectors
• wider bandwidths of the detectors

6
Shh of detectors
Conservative estimate
7
Target signals and observation time

• template search
• each detector applies an exchange threshold T on
  measured H
• exchanged data Dec 2004
• ALLEGRO data not yet available

8
exchanged thresholds
9
Detection efficiency

• AURIGA
• detection efficiency for ? signals
• measured by Monte Carlo injections of software
  signals

Exchange thresholds SNR4.5 for AURIGA SNR3.82
for EXPLORER and NAUTILUS
10
Detection efficiency for bursts
Maximum detection efficiency for transients with
flat Fourier amplitude at the detector
frequencies (?900 Hz)
Efficiency of the AURIGA ? matched filter for
Sine-Gaussian waveforms
11
Arrival time estimation

• AURIGA arrival time estimation for ? signals
• by Monte Carlo injections of software signals

• IGEC-2 is not yet able to measure light time
  delays among detectors

12
Exchanged candidate events

• amplitude histograms of exchanged events

13
Exchanged event rates

• Hourly rate of exchanged events

14
Self correlograms of exchanged events

• Histograms of the time lags among events of the
  same detector
• much more Poissonian
• than in IGEC-1

AU
50 seconds
EX
NA
15
Partial results duty cycle

• Coincidence periods with 3, 2 and single
  detectors
• with ALLEGRO data 3-fold configurations will
  dominate the observation time

16
Preliminary results accidental coincidences

• statistics of accidentals are in agreement with
  the model

counts
Poisson fit
2000 time shifts EX-NA pair
Number of accidental coincidences
17
Preliminary results accidental coincidences

• histograms of accidental coincidences as a
  function of the time shift
• ergodicity of accidentals vs time shift (same
  EX-NA)

counts
Color scale number of accidental
coincidences Dots mean Accidental coincidences
vs shift
Time shift s
18
cross correlograms of exchanged events

• Histograms of the time lags among all events from
  two different detectors
• Poisson model as in IGEC-1

AU-EX
AU-NA
EX-NA
19
Expected performances of IGEC-2 ?
Triple coincidences 106 time shifts, no
accidentals, 9.3 days false alarm rate lt 10-4 /
yr for Hgt1.4 10-21/Hz Double coincidences lower
false alarm rates than for IGEC-1
20
(No Transcript)