GEO600 Data Analysis Status

1
GEO600 Data Analysis Status

• Bernard Schutz
• Albert Einstein Institute
• Potsdam/Golm

2
GEO Participation in Science Runs

• GEO participated fully in S1, LSC analysis
  essentially finished. Two LSC working group
  co-chairs are GEO members MA Papa (pulsar
  searches), J Romano (stochastic).
• GEO did not participate in S2 installing signal
  recycling mirror.
• Just before beginning of S3, GEO scientists
  succeeded in learning how to control IFO with
  signal recycling mirror a crucial step in GW
  detector development. This demonstrates a
  technique that will be central to the Advanced
  LIGO upgrade.
• GEO participated in S3 for first week, then
  withdrew in order to improve sensitivity. Not yet
  clear when it will rejoin S3. Excellent
  stability duty cycle during first week was 95.
• See talks by J Smith, M Hewitson (Friday).

3
GEO Data Analysis People

• GEO coordinates its data activities within the
  project

• GEO Detector Characterization A Sintes, K Strain
• GEO Calibration H Ward, M Hewitson
• GEO data distribution and exchange S Grunewald,
  D Churches.

4
GW pulsar analysis

• See talks by G Woan, R Dupuis, A Vecchio
  (Thursday), poster by M A Papa.
• GEO has contributed the main pulsar search codes
  to the LSC (1) coherent demodulation, (2)
  Bayesian targeted searches using heterodyning,
  (3) Hough transform hierarchical search.
• First two codes used for analysis of S1 data from
  4 detectors LSC set 95 confidence upper limit
  ellipticity of PSR J 19392134 at 1284 Hz ? lt
  2.9 ? 10-4 (gr-qc/0308050, accepted PRD)
• S2 search widened to include all known pulsars
  plus first use of Hough transform to perform area
  search for unknown neutron stars.
• S2 analysis also includes searches for pulsars in
  binaries, searches for accreting neutron stars
  (Sco X-1).

5
Inspiral analysis

• See talk by T Cokelaer (Thursday)
• GEO generating PN templates, template bank. Tests
  of template placement using injected signals very
  satisfactory.

Detection statistic for S1 hardware injections Detection statistic for S1 hardware injections Detection statistic for S1 hardware injections Detection statistic for S1 hardware injections Detection statistic for S1 hardware injections Detection statistic for S1 hardware injections
Day injected candidates identified detected false alarms
250 135 128 126 93 2
251 167 169 167 100 1
252 92 85 81 88 4
Error in Chirp mass lt5, Total mass lt 10, Distance lt20 Error in Chirp mass lt5, Total mass lt 10, Distance lt20 Error in Chirp mass lt5, Total mass lt 10, Distance lt20 Error in Chirp mass lt5, Total mass lt 10, Distance lt20 Error in Chirp mass lt5, Total mass lt 10, Distance lt20 Error in Chirp mass lt5, Total mass lt 10, Distance lt20

• S3 plans spinning BH inspiral templates

6
Stochastic analysis

• See talks by J Romano, T Regimbau
• GEO-LIGO correlations inherently less sensitive
  due to longer baseline, but provide useful check
  on instrumental/analysis effects.
• First such correlations will be performed with S3
  data. Software for LAL/LALApps currently being
  validated.

7
Burst analysis

• See talks by S Heng (Thursday), R Balasubramanian
  (Saturday), poster by S Heng.
• The goal is to bring GEO S3 data into a
  multi-detector analysis. Different orientation
  and sensitivity bandwidth of GEO make this
  analysis interesting and non-trivial.
• Excess power and TFClusters generate events.
• Practical problems GEO data double-precision,
  GEO/HACR pipeline outside LAL.

8
GEO cluster computers

• Merlin (AEI/Golm) 180 dual-CPU Athlon nodes, 64
  TB disk space

• Explorer (Cardiff)80 dual-CPU Athlon nodes, 20
  TB disk space

• Tsunami (Birmingham) 100 dual-CPU Xeon nodes, 10
  TB disk space