RUN 07

You can access the following information:

Superchunks and results Chunks production and checks Data quality and segments Useful links and documentation Injections campaign Data and log files

MBTA standard offline search - O4a (May 2023 - Jan 2024)

This page documents the offline analysis of O4 Ligo and Virgo data to detect GW signals from coalescences of compact objects (black holes or neutron stars) using the MBTA code.
The data analysed in this page were collected during the O4a observation run (May 2023 - Jan 2024) by Ligo (Virgo being expected to join the run for O4b).
The data were split in several chunks of 2 weeks length, analysed separately. The official chunk limits are given in this page. You can find in the table below the links to the pages presenting the mbta offline analysis results for each chunk. At the bottom of the page, you also find details on how the quality of the data used for the analysis was defined and obtained: we use ANALYSIS-READY segments and CAT1 veto information from the Ligo and Virgo DetChar groups (according to the CBC group recommendations), on top of which MBTA applies its own gating.

List of superchunks and links to the results

Superchunk Date range (UTC) GPS time range Chunk length (days) Chunk length after CAT1 (days) Chunk length after gating (days) Results
≥ 2 ifo OK ≥ 1 ifo OK ≥ 2 ifo OK ≥ 1 ifo OK
01to04 2023-05-24 15:00
2023-07-18 10:33
1368975618
1373711624
54.81 31.05 (56.6 %) 48.08 (87.7 %) 30.98 (56.5 %) 48.06 (87.7 %)

READY

05to08 2023-07-18 10:33
2023-09-12 14:44
1373711624
1378565116
56.17 33.06 (58.9 %) 47.69 (84.9 %) 33.00 (58.7 %) 47.67 (84.9 %)

READY

09to12 2023-09-12 14:44
2023-11-07 15:03
1378565116
1383404656
56.01 33.60 (60.0 %) 48.54 (86.6 %) 33.55 (59.9 %) 48.52 (86.6 %)

READY

13to16 2023-11-07 15:03
2024-01-16 15:45
1383404656
1389455118
70.03 28.76 (41.1 %) 52.52 (75.0 %) 28.69 (41.0 %) 52.49 (75.0 %)

READY

Chunks production and checks

Link to individual chunk monitoring page Date range (UTC) GPS time range Chunk length (days) Chunk length after CAT1 (days) Production status
in at least Preprocessing Filtering Postfiltering
2 ifo 1 ifo DAT INJ02 INJ13 DAT INJ02 INJ13 DAT INJ02 INJ13
chunk01 2023-05-24 15:00
2023-06-06 14:30
1368975618
1370097052
12.98 5.66 11.77

OK

OK

OK

OK

OK

OK

OK

OK

OK

chunk02 2023-06-06 14:30
2023-06-20 14:21
1370097052
1371306087
13.99 7.55 12.19

OK

OK

OK

OK

OK

OK

OK

OK

OK

chunk03 2023-06-20 14:21
2023-07-04 22:47
1371306087
1372546081
14.35 8.75 12.55

OK

OK

OK

OK

OK

OK

OK

OK

OK

chunk04 2023-07-04 22:47
2023-07-18 10:33
1372546081
1373711624
13.49 9.08 11.57

OK

OK

OK

OK

OK

OK

OK

OK

OK

chunk05 2023-07-18 10:33
2023-08-01 14:44
1373711624
1374936316
14.17 7.71 11.22

OK

OK

OK

OK

OK

OK

OK

OK

OK

chunk06 2023-08-01 14:44
2023-08-15 14:25
1374936316
1376144770
13.99 6.65 11.17

OK

OK

OK

OK

OK

OK

OK

OK

OK

chunk07 2023-08-15 14:25
2023-08-29 14:44
1376144770
1377355516
14.01 9.97 12.70

OK

OK

OK

OK

OK

OK

OK

OK

OK

chunk08 2023-08-29 14:44
2023-09-12 14:44
1377355516
1378565116
14.00 8.74 12.59

OK

OK

OK

OK

OK

OK

OK

OK

OK

chunk09 2023-09-12 14:44
2023-09-26 14:44
1378565116
1379774716
14.00 9.63 12.57

OK

OK

OK

OK

OK

OK

OK

OK

OK

chunk10 2023-09-26 14:44
2023-10-10 14:44
1379774716
1380984316
14.00 9.13 12.40

OK

OK

OK

OK

OK

OK

OK

OK

OK

chunk11 2023-10-10 14:44
2023-10-24 15:04
1380984316
1382195103
14.01 8.19 12.25

OK

OK

OK

OK

OK

OK

OK

OK

OK

chunk12 2023-10-24 15:04
2023-11-07 15:03
1382195103
1383404656
14.00 6.66 11.32

OK

OK

OK

OK

OK

OK

OK

OK

OK

chunk13 2023-11-07 15:03
2023-11-21 07:59
1383404656
1384588801
13.71 7.15 11.74

OK

OK

OK

OK

OK

OK

OK

OK

OK

chunk14 2023-11-21 07:59
2023-12-05 14:46
1384588801
1385822808
14.28 5.84 9.38

OK

OK

OK

OK

OK

OK

OK

OK

OK

chunk15 2023-12-05 14:46
2023-12-21 15:15
1385822808
1387206946
16.02 5.15 12.85

OK

OK

OK

OK

OK

OK

OK

OK

OK

chunk16 2023-12-21 15:15
2024-01-16 15:45
1387206946
1389455118
26.02 10.61 18.55

OK

OK

OK

OK

OK

OK

OK

OK

OK

Data quality and segments

Some instructions to produce the list of good data quality segments are given here.
The list of good DQ segments is produced chunk by chunk and is available in /sps/virgo/USERS/mbta/O4/cat-flags/

1) The interferometer has to be in observing mode, also called DMT-ANALYSIS-READY in the segment database.
Recommendation of the ifo mode to be used for offline analysis is given in the ObsRun4 page.
The definitions of the ifo modes are given in the O4Flags page.
The DMT flags are defined in the dmt-runtime-configuration git page.

2) CAT1 flags are developped offline to reject additional bad data quality periods. The criteria are defined in a Veto Definer File (VDF) continuously updated during the run and can be found here.
The veto definer file for Burst and CBC analyses is now available for the full O4a period and tagged CAT1-HL-END_O4A-R2.

3) The good DQ segments and effective analysis times are given chunk by chunk in the links of the above table. They can also be found here for the whole O4a period, with the following summary numbers.

Three types of ascii files are produced: where IFO=H1, L1, H1L1 (for both H1 and L1 OK for analysis), H1orL1 (for at least H1 or L1 OK for analysis).
H1_ANALYSIS_READY-CAT1.txt and L1_ANALYSIS_READY-CAT1.txt are the files used in the offline analysis.

4) During MBTA preprocessing, h(t) strain data are gated for times outside the good DQ segments (i.e. a filter is applied to smoothly set h(t) to 0). MBTA also applies its own gating when the BNS range drops below 0.6 times the median range (the ranges are calculated for equal mass neutron stars of 1.4 Msol over the whole frequency range).
The good data quality segments after preprocessing for the whole O4a period (chunks 1 to 16) are available here for H1 and L1.
The other segment files for OR and AND combination of detectors are in this directory as well as a summary of the different periods.

chunk duration
(in days)
chunk duration after CAT1, in days chunk duration after gating, in days
(autogating loss in s)
HL good H good L good H or L good HL good H good L good H or L good
chunk01 12.98 5.66 (43.6 %) 6.19 (47.7 %) 11.24 (86.6 %) 11.77 (90.6 %) 5.65 (43.5 %)
(-1142.2 s)
6.18 (47.6 %)
(-765.1 s)
11.23 (86.5 %)
(-881.6 s)
11.76 (90.6 %)
(-504.5 s)
chunk02 13.99 7.55 (54.0 %) 8.45 (60.4 %) 11.29 (80.7 %) 12.19 (87.1 %) 7.53 (53.8 %)
(-1502.1 s)
8.44 (60.3 %)
(-961.6 s)
11.28 (80.6 %)
(-942.0 s)
12.19 (87.1 %)
(-401.5 s)
chunk03 14.35 8.75 (61.0 %) 9.57 (66.7 %) 11.74 (81.8 %) 12.55 (87.5 %) 8.74 (60.9 %)
(-1525.0 s)
9.55 (66.6 %)
(-984.1 s)
11.73 (81.7 %)
(-897.2 s)
12.55 (87.4 %)
(-356.2 s)
chunk04 13.49 9.08 (67.3 %) 9.52 (70.6 %) 11.13 (82.5 %) 11.57 (85.7 %) 9.06 (67.2 %)
(-1531.8 s)
9.51 (70.5 %)
(-959.5 s)
11.12 (82.4 %)
(-805.9 s)
11.56 (85.7 %)
(-233.7 s)
chunk05 14.17 7.71 (54.4 %) 9.93 (70.0 %) 9.00 (63.5 %) 11.22 (79.2 %) 7.69 (54.3 %)
(-1458.0 s)
9.91 (69.9 %)
(-1085.2 s)
9.00 (63.5 %)
(-739.5 s)
11.22 (79.1 %)
(-366.7 s)
chunk06 13.99 6.65 (47.5 %) 9.21 (65.9 %) 8.61 (61.5 %) 11.17 (79.9 %) 6.63 (47.4 %)
(-1232.8 s)
9.20 (65.8 %)
(-967.0 s)
8.60 (61.5 %)
(-681.5 s)
11.17 (79.8 %)
(-415.7 s)
chunk07 14.01 9.97 (71.1 %) 11.08 (79.1 %) 11.59 (82.7 %) 12.70 (90.7 %) 9.95 (71.0 %)
(-1527.1 s)
11.07 (79.0 %)
(-955.9 s)
11.58 (82.6 %)
(-834.7 s)
12.70 (90.6 %)
(-263.5 s)
chunk08 14.00 8.74 (62.4 %) 10.13 (72.3 %) 11.20 (80.0 %) 12.59 (89.9 %) 8.72 (62.3 %)
(-1328.4 s)
10.12 (72.3 %)
(-855.6 s)
11.19 (79.9 %)
(-754.4 s)
12.59 (89.9 %)
(-281.6 s)
chunk09 14.00 9.63 (68.8 %) 10.40 (74.3 %) 11.80 (84.3 %) 12.57 (89.8 %) 9.61 (68.7 %)
(-1294.9 s)
10.39 (74.2 %)
(-807.4 s)
11.79 (84.2 %)
(-725.4 s)
12.57 (89.8 %)
(-237.9 s)
chunk10 14.00 9.13 (65.2 %) 10.93 (78.1 %) 10.60 (75.7 %) 12.40 (88.5 %) 9.12 (65.1 %)
(-1266.1 s)
10.92 (78.0 %)
(-846.1 s)
10.59 (75.7 %)
(-689.5 s)
12.39 (88.5 %)
(-269.6 s)
chunk11 14.01 8.19 (58.4 %) 10.87 (77.6 %) 9.56 (68.2 %) 12.25 (87.4 %) 8.17 (58.3 %)
(-1166.3 s)
10.86 (77.5 %)
(-785.4 s)
9.56 (68.2 %)
(-670.6 s)
12.24 (87.4 %)
(-289.7 s)
chunk12 14.00 6.66 (47.5 %) 8.56 (61.1 %) 9.42 (67.3 %) 11.32 (80.9 %) 6.64 (47.5 %)
(-992.1 s)
8.55 (61.1 %)
(-744.7 s)
9.40 (67.2 %)
(-1001.9 s)
11.31 (80.8 %)
(-754.5 s)
chunk13 13.71 7.15 (52.2 %) 10.00 (72.9 %) 8.89 (64.9 %) 11.74 (85.6 %) 7.14 (52.1 %)
(-1190.3 s)
9.98 (72.9 %)
(-955.5 s)
8.89 (64.8 %)
(-573.1 s)
11.73 (85.6 %)
(-338.2 s)
chunk14 14.28 5.84 (40.9 %) 7.61 (53.3 %) 7.61 (53.3 %) 9.38 (65.7 %) 5.83 (40.8 %)
(-1056.0 s)
7.60 (53.2 %)
(-842.7 s)
7.60 (53.2 %)
(-578.2 s)
9.38 (65.6 %)
(-364.8 s)
chunk15 16.02 5.15 (32.2 %) 11.87 (74.1 %) 6.13 (38.3 %) 12.85 (80.2 %) 5.14 (32.1 %)
(-855.4 s)
11.86 (74.0 %)
(-964.1 s)
6.13 (38.2 %)
(-566.9 s)
12.85 (80.2 %)
(-675.6 s)
chunk16 26.02 10.61 (40.8 %) 15.51 (59.6 %) 13.65 (52.5 %) 18.55 (71.3 %) 10.58 (40.7 %)
(-2887.7 s)
15.48 (59.5 %)
(-2790.8 s)
13.64 (52.4 %)
(-1138.4 s)
18.54 (71.2 %)
(-1041.5 s)
Total 237.03 126.47 (53.4 %) 159.82 (67.4 %) 163.47 (69.0 %) 196.82 (83.0 %) 126.22 (53.2 %)
(-21956.2 s)
159.63 (67.3 %)
(-16270.7 s)
163.33 (68.9 %)
(-12480.8 s)
196.74 (83.0 %)
(-6795.2 s)

Documentation and useful links

The preprocessing was performed with FdIOServer version v8r61p0 (chunks 1 to 10) and version v8r62p1 (chunks 11 to 16).
Version v5r44 of MBTA was used for the filtering and postprocessing.

An MBTA documentation updated for its v5 version can be found here.
The documentation for the v4 version also contains some interesting details on the mbta processing.
FdIOServer is used for the preprocessing and the documentation of the Fd library can be consulted here.

Materials for the LVK internal review and signo-off for the O4 online and offline analysis are available on the O4ReviewMeetings wiki.

The production workflow, some instructions to produce the results and checks at all production steps are described in the mbta offline prod documentation.

Details about the injection set

The sets of injections are produced centrally by the Rate & Pop group for all the CBC pipelines.
inj distrib

The rate & pop injections have been produced in 4 trains, with a time interval of 24s between injections in each train. Each train is shifted by 6s with respect to the previous one. The injection h(t) are superposed to the h(t) data during the preprocessing step. Two runs are performed, one with the trains 0 and 2 spaced by 12s and another one with trains 1 and 3. Only train 3 contains BNS and NSBH events in addition to BBH. The number of injected events in one chunk is around 200 000 with a minority of NSBH (~4000) and BNS events (~5000).

Data files and logfiles

All output runs and logfiles are stored at the CCIN2P3 computing center.

The organisation of the folders is described here.

Preprocessing files (config, log and gwf):

/sps/virgo/USERS/mbta/O4/preprocessing/chunkXX/mbtaD/vYY/   (data)
/sps/virgo/USERS/mbta/O4/preprocessing/chunkXX/mbtaI/vYY/batchs-0-2/   (injection trains 0 and 2 spaced by 12s)
/sps/virgo/USERS/mbta/O4/preprocessing/chunkXX/mbtaI/vYY/batchs-1-3/   (injection trains 1 and 3 spaced by 12s)
where chunXX is the chunk number and vYY a production version number

Filtering files (config, log and gwf):

/sps/virgo/USERS/mbta/O4/results/chunkXX/mbtaD/runYY/   (data)
/sps/virgo/USERS/mbta/O4/results/chunkXX/mbtaI/runYY/batchs-0-2/   (injection trains 0 and 2 spaced by 12s)
/sps/virgo/USERS/mbta/O4/results/chunkXX/mbtaI/runYY/batchs-1-3/   (injection trains 1 and 3 spaced by 12s)
where chunXX is the chunk number and runYY a production version number

Postprocessing files (config, log and gwf):

/sps/virgo/USERS/mbta/O4/results/chunkXX/mbtaD/runYY/postfiltering/vZZ/   (data)
/sps/virgo/USERS/mbta/O4/results/chunkXX/mbtaI/runYY/batchs-0-2/postfiltering/vZZ/   (injection trains 0 and 2 spaced by 12s)
/sps/virgo/USERS/mbta/O4/results/chunkXX/mbtaI/runYY/batchs-1-3/postfiltering/vZZ/   (injection trains 1 and 3 spaced by 12s)
where chunXX is the chunk number, runYY the production version number of the filtering and vZZ the production version number of the postprocessing
These different paths with the version numbers used are given in each individual chunk web page. 

Files containing the analysis ready segments after CAT1 vetos:

/sps/virgo/USERS/mbta/O4/cat-flags/chunkXX/
where chunXX is the chunk number

The files containing the good data segments after mbta autogating are in the "segments" subdir of the preprocessing directories:

/sps/virgo/USERS/mbta/O4/preprocessing/chunkXX/mbtaD/vYY/segments/  (data)
/sps/virgo/USERS/mbta/O4/preprocessing/chunkXX/mbtaI/vYY/batchs-0-2/segments/   (injection trains 0 and 2 spaced by 12s)
/sps/virgo/USERS/mbta/O4/preprocessing/chunkXX/mbtaI/vYY/batchs-1-3/segments/   (injection trains 1 and 3 spaced by 12s)
where chunXX is the chunk number and vYY a production version number

The interferometer h(t) files are available on cvmfs or locally:

/sps/virgo/USERS/mbta/O4/HLVoffline/cvmfs/analysis-ready
/sps/virgo/USERS/mbta/O4/HLVoffline/local/analysis-ready

Injection files (characteristics of each individual injection event and h(t)):

/sps/virgo/USERS/mbta/O4/rpo4-offline

Template banks:

/sps/virgo/USERS/mbta/O4/banks/2023-02-01-preO4-HF-sorted-Q50-HF20ms

SNR Excess model:

sps/virgo/USERS/mbta/O4/snr-excess/2023-12-11-O4a-MainSearch/SNRexcessModel.gwf

pastro and FAR models are available in the superchunk folders:

/sps/virgo/USERS/mbta/O4/results/chunkXXtoYY/mbtaD/run07v01/FARofPastro.gwf