of 20
arXiv:1511.04398v2 [gr-qc] 18 Feb 2016
All-sky search for long-duration gravitational wave trans
ients with initial LIGO
B. P. Abbott
1
, R. Abbott
1
, T. D. Abbott
2
, M. R. Abernathy
1
, F. Acernese
3
,
4
, K. Ackley
5
, C. Adams
6
, T. Adams
7
,
P. Addesso
8
, R. X. Adhikari
1
, V. B. Adya
9
, C. Affeldt
9
, M. Agathos
10
, K. Agatsuma
10
, N. Aggarwal
11
,
O. D. Aguiar
12
, A. Ain
13
, P. Ajith
14
, B. Allen
9
,
15
,
16
, A. Allocca
17
,
18
, D. V. Amariutei
5
, S. B. Anderson
1
,
W. G. Anderson
15
, K. Arai
1
, M. C. Araya
1
, C. C. Arceneaux
19
, J. S. Areeda
20
, N. Arnaud
21
, K. G. Arun
22
,
G. Ashton
23
, M. Ast
24
, S. M. Aston
6
, P. Astone
25
, P. Aufmuth
16
, C. Aulbert
9
, S. Babak
26
, P. T. Baker
27
,
F. Baldaccini
28
,
29
, G. Ballardin
30
, S. W. Ballmer
31
, J. C. Barayoga
1
, S. E. Barclay
32
, B. C. Barish
1
, D. Barker
33
,
F. Barone
3
,
4
, B. Barr
32
, L. Barsotti
11
, M. Barsuglia
34
, D. Barta
35
, J. Bartlett
33
, I. Bartos
36
, R. Bassiri
37
,
A. Basti
17
,
18
, J. C. Batch
33
, C. Baune
9
, V. Bavigadda
30
, M. Bazzan
38
,
39
, B. Behnke
26
, M. Bejger
40
, C. Belczynski
41
,
A. S. Bell
32
, C. J. Bell
32
, B. K. Berger
1
, J. Bergman
33
, G. Bergmann
9
, C. P. L. Berry
42
, D. Bersanetti
43
,
44
,
A. Bertolini
10
, J. Betzwieser
6
, S. Bhagwat
31
, R. Bhandare
45
, I. A. Bilenko
46
, G. Billingsley
1
, J. Birch
6
, R. Birney
47
,
S. Biscans
11
, A. Bisht
9
,
16
, M. Bitossi
30
, C. Biwer
31
, M. A. Bizouard
21
, J. K. Blackburn
1
, C. D. Blair
48
, D. Blair
48
,
R. M. Blair
33
, S. Bloemen
10
,
49
, O. Bock
9
, T. P. Bodiya
11
, M. Boer
50
, G. Bogaert
50
, C. Bogan
9
, A. Bohe
26
,
P. Bojtos
51
, C. Bond
42
, F. Bondu
52
, R. Bonnand
7
, R. Bork
1
, V. Boschi
18
,
17
, S. Bose
53
,
13
, A. Bozzi
30
,
C. Bradaschia
18
, P. R. Brady
15
, V. B. Braginsky
46
, M. Branchesi
54
,
55
, J. E. Brau
56
, T. Briant
57
, A. Brillet
50
,
M. Brinkmann
9
, V. Brisson
21
, P. Brockill
15
, A. F. Brooks
1
, D. A. Brown
31
, D. Brown
5
, D. D. Brown
42
,
N. M. Brown
11
, C. C. Buchanan
2
, A. Buikema
11
, T. Bulik
41
, H. J. Bulten
58
,
10
, A. Buonanno
26
,
59
, D. Buskulic
7
,
C. Buy
34
, R. L. Byer
37
, L. Cadonati
60
, G. Cagnoli
61
, C. Cahillane
1
, J. Calder ́on Bustillo
62
,
60
, T. Callister
1
,
E. Calloni
63
,
4
, J. B. Camp
64
, K. C. Cannon
65
, J. Cao
66
, C. D. Capano
9
, E. Capocasa
34
, F. Carbognani
30
,
S. Caride
67
, J. Casanueva Diaz
21
, C. Casentini
68
,
69
, S. Caudill
15
, M. Cavagli`a
19
, F. Cavalier
21
, R. Cavalieri
30
,
G. Cella
18
, C. Cepeda
1
, L. Cerboni Baiardi
54
,
55
, G. Cerretani
17
,
18
, E. Cesarini
68
,
69
, R. Chakraborty
1
,
T. Chalermsongsak
1
, S. J. Chamberlin
15
, M. Chan
32
, S. Chao
70
, P. Charlton
71
, E. Chassande-Mottin
34
,
H. Y. Chen
72
, Y. Chen
73
, C. Cheng
70
, A. Chincarini
44
, A. Chiummo
30
, H. S. Cho
74
, M. Cho
59
, J. H. Chow
75
,
N. Christensen
76
, Q. Chu
48
, S. Chua
57
, S. Chung
48
, G. Ciani
5
, F. Clara
33
, J. A. Clark
60
, F. Cleva
50
, E. Coccia
68
,
77
,
P.-F. Cohadon
57
, A. Colla
78
,
25
, C. G. Collette
79
, M. Constancio Jr.
12
, A. Conte
78
,
25
, L. Conti
39
, D. Cook
33
,
T. R. Corbitt
2
, N. Cornish
27
, A. Corsi
80
, S. Cortese
30
, C. A. Costa
12
, M. W. Coughlin
76
, S. B. Coughlin
81
,
J.-P. Coulon
50
, S. T. Countryman
36
, P. Couvares
1
, D. M. Coward
48
, M. J. Cowart
6
, D. C. Coyne
1
, R. Coyne
80
,
K. Craig
32
, J. D. E. Creighton
15
, J. Cripe
2
, S. G. Crowder
82
, A. Cumming
32
, L. Cunningham
32
, E. Cuoco
30
,
T. Dal Canton
9
, S. L. Danilishin
32
, S. D’Antonio
69
, K. Danzmann
16
,
9
, N. S. Darman
83
, V. Dattilo
30
, I. Dave
45
,
H. P. Daveloza
84
, M. Davier
21
, G. S. Davies
32
, E. J. Daw
85
, R. Day
30
, D. DeBra
37
, G. Debreczeni
35
, J. Degallaix
61
,
M. De Laurentis
63
,
4
, S. Del ́eglise
57
, W. Del Pozzo
42
, T. Denker
9
,
16
, T. Dent
9
, H. Dereli
50
, V. Dergachev
1
,
R. DeRosa
6
, R. De Rosa
63
,
4
, R. DeSalvo
8
, S. Dhurandhar
13
, M. C. D ́ıaz
84
, L. Di Fiore
4
, M. Di Giovanni
78
,
25
,
A. Di Lieto
17
,
18
, I. Di Palma
26
,
9
, A. Di Virgilio
18
, G. Dojcinoski
86
, V. Dolique
61
, F. Donovan
11
, K. L. Dooley
19
,
S. Doravari
6
, R. Douglas
32
, T. P. Downes
15
, M. Drago
9
,
87
,
88
, R. W. P. Drever
1
, J. C. Driggers
33
, Z. Du
66
,
M. Ducrot
7
, S. E. Dwyer
33
, T. B. Edo
85
, M. C. Edwards
76
, A. Effler
6
, H.-B. Eggenstein
9
, P. Ehrens
1
,
J. M. Eichholz
5
, S. S. Eikenberry
5
, W. Engels
73
, R. C. Essick
11
, T. Etzel
1
, M. Evans
11
, T. M. Evans
6
, R. Everett
89
,
M. Factourovich
36
, V. Fafone
68
,
69
,
77
, H. Fair
31
, S. Fairhurst
81
, X. Fan
66
, Q. Fang
48
, S. Farinon
44
, B. Farr
72
,
W. M. Farr
42
, M. Favata
86
, M. Fays
81
, H. Fehrmann
9
, M. M. Fejer
37
, I. Ferrante
17
,
18
, E. C. Ferreira
12
, F. Ferrini
30
,
F. Fidecaro
17
,
18
, I. Fiori
30
, R. P. Fisher
31
, R. Flaminio
61
, M. Fletcher
32
, J.-D. Fournier
50
, S. Franco
21
, S. Frasca
78
,
25
,
F. Frasconi
18
, Z. Frei
51
, A. Freise
42
, R. Frey
56
, V. Frey
21
, T. T. Fricke
9
, P. Fritschel
11
, V. V. Frolov
6
, P. Fulda
5
,
M. Fyffe
6
, H. A. G. Gabbard
19
, J. R. Gair
90
, L. Gammaitoni
28
,
29
, S. G. Gaonkar
13
, F. Garufi
63
,
4
, A. Gatto
34
,
G. Gaur
91
,
92
, N. Gehrels
64
, G. Gemme
44
, B. Gendre
50
, E. Genin
30
, A. Gennai
18
, J. George
45
, L. Gergely
93
,
V. Germain
7
, A. Ghosh
14
, S. Ghosh
10
,
49
, J. A. Giaime
2
,
6
, K. D. Giardina
6
, A. Giazotto
18
, K. Gill
94
, A. Glaefke
32
,
E. Goetz
67
, R. Goetz
5
, L. Gondan
51
, G. Gonz ́alez
2
, J. M. Gonzalez Castro
17
,
18
, A. Gopakumar
95
, N. A. Gordon
32
,
M. L. Gorodetsky
46
, S. E. Gossan
1
, M. Gosselin
30
, R. Gouaty
7
, C. Graef
32
, P. B. Graff
64
,
59
, M. Granata
61
,
A. Grant
32
, S. Gras
11
, C. Gray
33
, G. Greco
54
,
55
, A. C. Green
42
, P. Groot
49
, H. Grote
9
, S. Grunewald
26
,
G. M. Guidi
54
,
55
, X. Guo
66
, A. Gupta
13
, M. K. Gupta
92
, K. E. Gushwa
1
, E. K. Gustafson
1
, R. Gustafson
67
,
J. J. Hacker
20
, B. R. Hall
53
, E. D. Hall
1
, G. Hammond
32
, M. Haney
95
, M. M. Hanke
9
, J. Hanks
33
, C. Hanna
89
,
M. D. Hannam
81
, J. Hanson
6
, T. Hardwick
2
, J. Harms
54
,
55
, G. M. Harry
96
, I. W. Harry
26
, M. J. Hart
32
,
M. T. Hartman
5
, C.-J. Haster
42
, K. Haughian
32
, A. Heidmann
57
, M. C. Heintze
5
,
6
, H. Heitmann
50
, P. Hello
21
,
G. Hemming
30
, M. Hendry
32
, I. S. Heng
32
, J. Hennig
32
, A. W. Heptonstall
1
, M. Heurs
9
,
16
, S. Hild
32
, D. Hoak
97
,
K. A. Hodge
1
, D. Hofman
61
, S. E. Hollitt
98
, K. Holt
6
, D. E. Holz
72
, P. Hopkins
81
, D. J. Hosken
98
, J. Hough
32
,
E. A. Houston
32
, E. J. Howell
48
, Y. M. Hu
32
, S. Huang
70
, E. A. Huerta
99
, D. Huet
21
, B. Hughey
94
, S. Husa
62
,
2
S. H. Huttner
32
, T. Huynh-Dinh
6
, A. Idrisy
89
, N. Indik
9
, D. R. Ingram
33
, R. Inta
80
, H. N. Isa
32
, J.-M. Isac
57
,
M. Isi
1
, G. Islas
20
, T. Isogai
11
, B. R. Iyer
14
, K. Izumi
33
, T. Jacqmin
57
, H. Jang
74
, K. Jani
60
, P. Jaranowski
100
,
S. Jawahar
101
, F. Jim ́enez-Forteza
62
, W. W. Johnson
2
, D. I. Jones
23
, R. Jones
32
, R.J.G. Jonker
10
, L. Ju
48
,
Haris K
102
, C. V. Kalaghatgi
22
, V. Kalogera
103
, S. Kandhasamy
19
, G. Kang
74
, J. B. Kanner
1
, S. Karki
56
,
M. Kasprzack
2
,
21
,
30
, E. Katsavounidis
11
, W. Katzman
6
, S. Kaufer
16
, T. Kaur
48
, K. Kawabe
33
, F. Kawazoe
9
,
F. K ́ef ́elian
50
, M. S. Kehl
65
, D. Keitel
9
, D. B. Kelley
31
, W. Kells
1
, R. Kennedy
85
, J. S. Key
84
, A. Khalaidovski
9
,
F. Y. Khalili
46
, S. Khan
81
, Z. Khan
92
, E. A. Khazanov
104
, N. Kijbunchoo
33
, C. Kim
74
, J. Kim
105
, K. Kim
106
,
N. Kim
74
, N. Kim
37
, Y.-M. Kim
105
, E. J. King
98
, P. J. King
33
, D. L. Kinzel
6
, J. S. Kissel
33
, L. Kleybolte
24
,
S. Klimenko
5
, S. M. Koehlenbeck
9
, K. Kokeyama
2
, S. Koley
10
, V. Kondrashov
1
, A. Kontos
11
, M. Korobko
24
,
W. Z. Korth
1
, I. Kowalska
41
, D. B. Kozak
1
, V. Kringel
9
, B. Krishnan
9
, A. Kr ́olak
107
,
108
, C. Krueger
16
, G. Kuehn
9
,
P. Kumar
65
, L. Kuo
70
, A. Kutynia
107
, B. D. Lackey
31
, M. Landry
33
, J. Lange
109
, B. Lantz
37
, P. D. Lasky
110
,
A. Lazzarini
1
, C. Lazzaro
60
,
39
, P. Leaci
26
,
78
,
25
, S. Leavey
32
, E. Lebigot
34
,
66
, C. H. Lee
105
, H. K. Lee
106
,
H. M. Lee
111
, K. Lee
32
, M. Leonardi
87
,
88
, J. R. Leong
9
, N. Leroy
21
, N. Letendre
7
, Y. Levin
110
, B. M. Levine
33
,
T. G. F. Li
1
, A. Libson
11
, T. B. Littenberg
103
, N. A. Lockerbie
101
, J. Logue
32
, A. L. Lombardi
97
, J. E. Lord
31
,
M. Lorenzini
77
, V. Loriette
112
, M. Lormand
6
, G. Losurdo
55
, J. D. Lough
9
,
16
, H. L ̈uck
16
,
9
, A. P. Lundgren
9
, J. Luo
76
,
R. Lynch
11
, Y. Ma
48
, T. MacDonald
37
, B. Machenschalk
9
, M. MacInnis
11
, D. M. Macleod
2
, F. Maga ̃na-Sandoval
31
,
R. M. Magee
53
, M. Mageswaran
1
, E. Majorana
25
, I. Maksimovic
112
, V. Malvezzi
68
,
69
, N. Man
50
, I. Mandel
42
,
V. Mandic
82
, V. Mangano
78
,
25
,
32
, G. L. Mansell
75
, M. Manske
15
, M. Mantovani
30
, F. Marchesoni
113
,
29
, F. Marion
7
,
S. M ́arka
36
, Z. M ́arka
36
, A. S. Markosyan
37
, E. Maros
1
, F. Martelli
54
,
55
, L. Martellini
50
, I. W. Martin
32
,
R. M. Martin
5
, D. V. Martynov
1
, J. N. Marx
1
, K. Mason
11
, A. Masserot
7
, T. J. Massinger
31
, M. Masso-Reid
32
,
F. Matichard
11
, L. Matone
36
, N. Mavalvala
11
, N. Mazumder
53
, G. Mazzolo
9
, R. McCarthy
33
, D. E. McClelland
75
,
S. McCormick
6
, S. C. McGuire
114
, G. McIntyre
1
, J. McIver
97
, S. T. McWilliams
99
, D. Meacher
50
,
G. D. Meadors
26
,
9
, J. Meidam
10
, A. Melatos
83
, G. Mendell
33
, D. Mendoza-Gandara
9
, R. A. Mercer
15
,
M. Merzougui
50
, S. Meshkov
1
, C. Messenger
32
, C. Messick
89
, P. M. Meyers
82
, F. Mezzani
25
,
78
, H. Miao
42
,
C. Michel
61
, H. Middleton
42
, E. E. Mikhailov
115
, L. Milano
63
,
4
, J. Miller
11
, M. Millhouse
27
, Y. Minenkov
69
,
J. Ming
26
,
9
, S. Mirshekari
116
, C. Mishra
14
, S. Mitra
13
, V. P. Mitrofanov
46
, G. Mitselmakher
5
, R. Mittleman
11
,
A. Moggi
18
, S. R. P. Mohapatra
11
, M. Montani
54
,
55
, B. C. Moore
86
, C. J. Moore
90
, D. Moraru
33
, G. Moreno
33
,
S. R. Morriss
84
, K. Mossavi
9
, B. Mours
7
, C. M. Mow-Lowry
42
, C. L. Mueller
5
, G. Mueller
5
, A. W. Muir
81
,
Arunava Mukherjee
14
, D. Mukherjee
15
, S. Mukherjee
84
, A. Mullavey
6
, J. Munch
98
, D. J. Murphy
36
, P. G. Murray
32
,
A. Mytidis
5
, I. Nardecchia
68
,
69
, L. Naticchioni
78
,
25
, R. K. Nayak
117
, V. Necula
5
, K. Nedkova
97
, G. Nelemans
10
,
49
,
M. Neri
43
,
44
, A. Neunzert
67
, G. Newton
32
, T. T. Nguyen
75
, A. B. Nielsen
9
, S. Nissanke
49
,
10
, A. Nitz
31
, F. Nocera
30
,
D. Nolting
6
, M. E. N. Normandin
84
, L. K. Nuttall
31
, J. Oberling
33
, E. Ochsner
15
, J. O’Dell
118
, E. Oelker
11
,
G. H. Ogin
119
, J. J. Oh
120
, S. H. Oh
120
, F. Ohme
81
, M. Oliver
62
, P. Oppermann
9
, Richard J. Oram
6
, B. O’Reilly
6
,
R. O’Shaughnessy
109
, C. D. Ott
73
, D. J. Ottaway
98
, R. S. Ottens
5
, H. Overmier
6
, B. J. Owen
80
, A. Pai
102
,
S. A. Pai
45
, J. R. Palamos
56
, O. Palashov
104
, C. Palomba
25
, A. Pal-Singh
24
, H. Pan
70
, C. Pankow
15
,
103
,
F. Pannarale
81
, B. C. Pant
45
, F. Paoletti
30
,
18
, A. Paoli
30
, M. A. Papa
26
,
15
,
9
, H. R. Paris
37
, W. Parker
6
,
D. Pascucci
32
, A. Pasqualetti
30
, R. Passaquieti
17
,
18
, D. Passuello
18
, Z. Patrick
37
, B. L. Pearlstone
32
, M. Pedraza
1
,
R. Pedurand
61
, L. Pekowsky
31
, A. Pele
6
, S. Penn
121
, R. Pereira
36
, A. Perreca
1
, M. Phelps
32
, O. Piccinni
78
,
25
,
M. Pichot
50
, F. Piergiovanni
54
,
55
, V. Pierro
8
, G. Pillant
30
, L. Pinard
61
, I. M. Pinto
8
, M. Pitkin
32
,
R. Poggiani
17
,
18
, A. Post
9
, J. Powell
32
, J. Prasad
13
, V. Predoi
81
, S. S. Premachandra
110
, T. Prestegard
82
,
L. R. Price
1
, M. Prijatelj
30
, M. Principe
8
, S. Privitera
26
, G. A. Prodi
87
,
88
, L. Prokhorov
46
, M. Punturo
29
,
P. Puppo
25
, M. P ̈urrer
81
, H. Qi
15
, J. Qin
48
, V. Quetschke
84
, E. A. Quintero
1
, R. Quitzow-James
56
, F. J. Raab
33
,
D. S. Rabeling
75
, H. Radkins
33
, P. Raffai
51
, S. Raja
45
, M. Rakhmanov
84
, P. Rapagnani
78
,
25
, V. Raymond
26
,
M. Razzano
17
,
18
, V. Re
68
,
69
, J. Read
20
, C. M. Reed
33
, T. Regimbau
50
, L. Rei
44
, S. Reid
47
, D. H. Reitze
1
,
5
,
H. Rew
115
, F. Ricci
78
,
25
, K. Riles
67
, N. A. Robertson
1
,
32
, R. Robie
32
, F. Robinet
21
, A. Rocchi
69
, L. Rolland
7
,
J. G. Rollins
1
, V. J. Roma
56
, J. D. Romano
84
, R. Romano
3
,
4
, G. Romanov
115
, J. H. Romie
6
, D. Rosi ́nska
122
,
40
,
S. Rowan
32
, A. R ̈udiger
9
, P. Ruggi
30
, K. Ryan
33
, S. Sachdev
1
, T. Sadecki
33
, L. Sadeghian
15
, M. Saleem
102
,
F. Salemi
9
, A. Samajdar
117
, L. Sammut
83
, E. J. Sanchez
1
, V. Sandberg
33
, B. Sandeen
103
, J. R. Sanders
67
,
B. Sassolas
61
, P. R. Saulson
31
, O. Sauter
67
, R. Savage
33
, A. Sawadsky
16
, P. Schale
56
, R. Schilling
,
9
, J. Schmidt
9
,
P. Schmidt
1
,
73
, R. Schnabel
24
, R. M. S. Schofield
56
, A. Sch ̈onbeck
24
, E. Schreiber
9
, D. Schuette
9
,
16
, B. F. Schutz
81
,
J. Scott
32
, S. M. Scott
75
, D. Sellers
6
, D. Sentenac
30
, V. Sequino
68
,
69
, A. Sergeev
104
, G. Serna
20
, Y. Setyawati
49
,
10
,
A. Sevigny
33
, D. A. Shaddock
75
, S. Shah
10
,
49
, M. S. Shahriar
103
, M. Shaltev
9
, Z. Shao
1
, B. Shapiro
37
, P. Shawhan
59
,
A. Sheperd
15
, D. H. Shoemaker
11
, D. M. Shoemaker
60
, K. Siellez
50
, X. Siemens
15
, D. Sigg
33
, A. D. Silva
12
,
3
D. Simakov
9
, A. Singer
1
, L. P. Singer
64
, A. Singh
26
,
9
, R. Singh
2
, A. M. Sintes
62
, B. J. J. Slagmolen
75
, J. R. Smith
20
,
N. D. Smith
1
, R. J. E. Smith
1
, E. J. Son
120
, B. Sorazu
32
, F. Sorrentino
44
, T. Souradeep
13
, A. K. Srivastava
92
,
A. Staley
36
, M. Steinke
9
, J. Steinlechner
32
, S. Steinlechner
32
, D. Steinmeyer
9
,
16
, B. C. Stephens
15
, R. Stone
84
,
K. A. Strain
32
, N. Straniero
61
, G. Stratta
54
,
55
, N. A. Strauss
76
, S. Strigin
46
, R. Sturani
116
, A. L. Stuver
6
,
T. Z. Summerscales
123
, L. Sun
83
, P. J. Sutton
81
, B. L. Swinkels
30
, M. J. Szczepanczyk
94
, M. Tacca
34
, D. Talukder
56
,
D. B. Tanner
5
, M. T ́apai
93
, S. P. Tarabrin
9
, A. Taracchini
26
, R. Taylor
1
, T. Theeg
9
, M. P. Thirugnanasambandam
1
,
E. G. Thomas
42
, M. Thomas
6
, P. Thomas
33
, K. A. Thorne
6
, K. S. Thorne
73
, E. Thrane
110
, S. Tiwari
77
,
V. Tiwari
81
, C. Tomlinson
85
, M. Tonelli
17
,
18
, C. V. Torres
,
84
, C. I. Torrie
1
, D. T ̈oyr ̈a
42
, F. Travasso
28
,
29
,
G. Traylor
6
, D. Trifir`o
19
, M. C. Tringali
87
,
88
, L. Trozzo
124
,
18
, M. Tse
11
, M. Turconi
50
, D. Tuyenbayev
84
,
D. Ugolini
125
, C. S. Unnikrishnan
95
, A. L. Urban
15
, S. A. Usman
31
, H. Vahlbruch
16
, G. Vajente
1
, G. Valdes
84
,
N. van Bakel
10
, M. van Beuzekom
10
, J. F. J. van den Brand
58
,
10
, C. van den Broeck
10
, L. van der Schaaf
10
,
M. V. van der Sluys
10
,
49
, J. V. van Heijningen
10
, A. A. van Veggel
32
, M. Vardaro
38
,
39
, S. Vass
1
, M. Vas ́uth
35
,
R. Vaulin
11
, A. Vecchio
42
, G. Vedovato
39
, J. Veitch
42
, P. J. Veitch
98
, K. Venkateswara
126
, D. Verkindt
7
,
F. Vetrano
54
,
55
, A. Vicer ́e
54
,
55
, S. Vinciguerra
42
, J.-Y. Vinet
50
, S. Vitale
11
, T. Vo
31
, H. Vocca
28
,
29
, C. Vorvick
33
,
W. D. Vousden
42
, S. P. Vyatchanin
46
, A. R. Wade
75
, L. E. Wade
15
, M. Wade
15
, M. Walker
2
, L. Wallace
1
,
S. Walsh
15
, G. Wang
77
, H. Wang
42
, M. Wang
42
, X. Wang
66
, Y. Wang
48
, R. L. Ward
75
, J. Warner
33
, M. Was
7
,
B. Weaver
33
, L.-W. Wei
50
, M. Weinert
9
, A. J. Weinstein
1
, R. Weiss
11
, T. Welborn
6
, L. Wen
48
, P. Weßels
9
,
T. Westphal
9
, K. Wette
9
, J. T. Whelan
109
,
9
, D. J. White
85
, B. F. Whiting
5
, R. D. Williams
1
, A. R. Williamson
81
,
J. L. Willis
127
, B. Willke
16
,
9
, M. H. Wimmer
9
,
16
, W. Winkler
9
, C. C. Wipf
1
, H. Wittel
9
,
16
, G. Woan
32
, J. Worden
33
,
J. L. Wright
32
, G. Wu
6
, J. Yablon
103
, W. Yam
11
, H. Yamamoto
1
, C. C. Yancey
59
, H. Yu
11
, M. Yvert
7
,
A. Zadro ̇zny
107
, L. Zangrando
39
, M. Zanolin
94
, J.-P. Zendri
39
, M. Zevin
103
, F. Zhang
11
, L. Zhang
1
, M. Zhang
115
,
Y. Zhang
109
, C. Zhao
48
, M. Zhou
103
, Z. Zhou
103
, X. J. Zhu
48
, M. E. Zucker
11
, S. E. Zuraw
97
, and J. Zweizig
1
(LIGO Scientific Collaboration and Virgo Collaboration)
Deceased, May 2015.
Deceased, March 2015.
1
LIGO, California Institute of Technology, Pasadena, CA 911
25, USA
2
Louisiana State University, Baton Rouge, LA 70803, USA
3
Universit`a di Salerno, Fisciano, I-84084 Salerno, Italy
4
INFN, Sezione di Napoli, Complesso Universitario di Monte S
.Angelo, I-80126 Napoli, Italy
5
University of Florida, Gainesville, FL 32611, USA
6
LIGO Livingston Observatory, Livingston, LA 70754, USA
7
Laboratoire d’Annecy-le-Vieux de Physique des Particules
(LAPP),
Universit ́e Savoie Mont Blanc, CNRS/IN2P3, F-74941 Annecy
-le-Vieux, France
8
University of Sannio at Benevento, I-82100 Benevento,
Italy and INFN, Sezione di Napoli, I-80100 Napoli, Italy
9
Albert-Einstein-Institut, Max-Planck-Institut f ̈ur Gra
vitationsphysik, D-30167 Hannover, Germany
10
Nikhef, Science Park, 1098 XG Amsterdam, The Netherlands
11
LIGO, Massachusetts Institute of Technology, Cambridge, M
A 02139, USA
12
Instituto Nacional de Pesquisas Espaciais, 12227-010 S ̃ao
Jos ́e dos Campos, SP, Brazil
13
Inter-University Centre for Astronomy and Astrophysics, P
une 411007, India
14
International Centre for Theoretical Sciences, Tata Insti
tute of Fundamental Research, Bangalore 560012, India
15
University of Wisconsin-Milwaukee, Milwaukee, WI 53201, U
SA
16
Leibniz Universit ̈at Hannover, D-30167 Hannover, Germany
17
Universit`a di Pisa, I-56127 Pisa, Italy
18
INFN, Sezione di Pisa, I-56127 Pisa, Italy
19
The University of Mississippi, University, MS 38677, USA
20
California State University Fullerton, Fullerton, CA 9283
1, USA
21
LAL, Univ Paris-Sud, CNRS/IN2P3, Orsay, France
22
Chennai Mathematical Institute, Chennai, India
23
University of Southampton, Southampton SO17 1BJ, United Ki
ngdom
24
Universit ̈at Hamburg, D-22761 Hamburg, Germany
25
INFN, Sezione di Roma, I-00185 Roma, Italy
26
Albert-Einstein-Institut, Max-Planck-Institut f ̈ur Gra
vitationsphysik, D-14476 Potsdam-Golm, Germany
27
Montana State University, Bozeman, MT 59717, USA
28
Universit`a di Perugia, I-06123 Perugia, Italy
29
INFN, Sezione di Perugia, I-06123 Perugia, Italy
30
European Gravitational Observatory (EGO), I-56021 Cascin
a, Pisa, Italy
31
Syracuse University, Syracuse, NY 13244, USA
32
SUPA, University of Glasgow, Glasgow G12 8QQ, United Kingdo
m
33
LIGO Hanford Observatory, Richland, WA 99352, USA
4
34
APC, AstroParticule et Cosmologie, Universit ́e Paris Dide
rot,
CNRS/IN2P3, CEA/Irfu, Observatoire de Paris,
Sorbonne Paris Cit ́e, F-75205 Paris Cedex 13, France
35
Wigner RCP, RMKI, H-1121 Budapest, Konkoly Thege Mikl ́os ́u
t 29-33, Hungary
36
Columbia University, New York, NY 10027, USA
37
Stanford University, Stanford, CA 94305, USA
38
Universit`a di Padova, Dipartimento di Fisica e Astronomia
, I-35131 Padova, Italy
39
INFN, Sezione di Padova, I-35131 Padova, Italy
40
CAMK-PAN, 00-716 Warsaw, Poland
41
Astronomical Observatory Warsaw University, 00-478 Warsa
w, Poland
42
University of Birmingham, Birmingham B15 2TT, United Kingd
om
43
Universit`a degli Studi di Genova, I-16146 Genova, Italy
44
INFN, Sezione di Genova, I-16146 Genova, Italy
45
RRCAT, Indore MP 452013, India
46
Faculty of Physics, Lomonosov Moscow State University, Mos
cow 119991, Russia
47
SUPA, University of the West of Scotland, Paisley PA1 2BE, Un
ited Kingdom
48
University of Western Australia, Crawley, Western Austral
ia 6009, Australia
49
Department of Astrophysics/IMAPP, Radboud University Nij
megen,
P.O. Box 9010, 6500 GL Nijmegen, The Netherlands
50
ARTEMIS, Universit ́e Cˆote d’Azur, CNRS and Observatoire d
e la Cˆote d’Azur, F-06304 Nice, France
51
MTA E ̈otv ̈os University, “Lendulet” Astrophysics Researc
h Group, Budapest 1117, Hungary
52
Institut de Physique de Rennes, CNRS, Universit ́e de Rennes
1, F-35042 Rennes, France
53
Washington State University, Pullman, WA 99164, USA
54
Universit`a degli Studi di Urbino ’Carlo Bo’, I-61029 Urbin
o, Italy
55
INFN, Sezione di Firenze, I-50019 Sesto Fiorentino, Firenz
e, Italy
56
University of Oregon, Eugene, OR 97403, USA
57
Laboratoire Kastler Brossel, UPMC-Sorbonne Universit ́es
, CNRS,
ENS-PSL Research University, Coll`ege de France, F-75005 P
aris, France
58
VU University Amsterdam, 1081 HV Amsterdam, The Netherland
s
59
University of Maryland, College Park, MD 20742, USA
60
Center for Relativistic Astrophysics and School of Physics
,
Georgia Institute of Technology, Atlanta, GA 30332, USA
61
Laboratoire des Mat ́eriaux Avanc ́es (LMA), IN2P3/CNRS,
Universit ́e de Lyon, F-69622 Villeurbanne, Lyon, France
62
Universitat de les Illes Balears—IEEC, E-07122 Palma de Mal
lorca, Spain
63
Universit`a di Napoli ’Federico II’, Complesso Universita
rio di Monte S.Angelo, I-80126 Napoli, Italy
64
NASA/Goddard Space Flight Center, Greenbelt, MD 20771, USA
65
Canadian Institute for Theoretical Astrophysics,
University of Toronto, Toronto, Ontario M5S 3H8, Canada
66
Tsinghua University, Beijing 100084, China
67
University of Michigan, Ann Arbor, MI 48109, USA
68
Universit`a di Roma Tor Vergata, I-00133 Roma, Italy
69
INFN, Sezione di Roma Tor Vergata, I-00133 Roma, Italy
70
National Tsing Hua University, Hsinchu City, Taiwan 30013,
R.O.C.
71
Charles Sturt University, Wagga Wagga, New South Wales 2678
, Australia
72
University of Chicago, Chicago, IL 60637, USA
73
Caltech CaRT, Pasadena, CA 91125, USA
74
Korea Institute of Science and Technology Information, Dae
jeon 305-806, Korea
75
Australian National University, Canberra, Australian Cap
ital Territory 0200, Australia
76
Carleton College, Northfield, MN 55057, USA
77
INFN, Gran Sasso Science Institute, I-67100 L’Aquila, Ital
y
78
Universit`a di Roma ’La Sapienza’, I-00185 Roma, Italy
79
University of Brussels, Brussels 1050, Belgium
80
Texas Tech University, Lubbock, TX 79409, USA
81
Cardiff University, Cardiff CF24 3AA, United Kingdom
82
University of Minnesota, Minneapolis, MN 55455, USA
83
The University of Melbourne, Parkville, Victoria 3010, Aus
tralia
84
The University of Texas Rio Grande Valley, Brownsville, TX 7
8520, USA
85
The University of Sheffield, Sheffield S10 2TN, United Kingdom
86
Montclair State University, Montclair, NJ 07043, USA
87
Universit`a di Trento, Dipartimento di Fisica, I-38123 Pov
o, Trento, Italy
88
INFN, Trento Institute for Fundamental Physics and Applica
tions, I-38123 Povo, Trento, Italy
89
The Pennsylvania State University, University Park, PA 168
02, USA
90
University of Cambridge, Cambridge CB2 1TN, United Kingdom
5
91
Indian Institute of Technology, Gandhinagar Ahmedabad Guj
arat 382424, India
92
Institute for Plasma Research, Bhat, Gandhinagar 382428, I
ndia
93
University of Szeged, D ́om t ́er 9, Szeged 6720, Hungary
94
Embry-Riddle Aeronautical University, Prescott, AZ 86301
, USA
95
Tata Institute for Fundamental Research, Mumbai 400005, In
dia
96
American University, Washington, D.C. 20016, USA
97
University of Massachusetts-Amherst, Amherst, MA 01003, U
SA
98
University of Adelaide, Adelaide, South Australia 5005, Au
stralia
99
West Virginia University, Morgantown, WV 26506, USA
100
University of Bia lystok, 15-424 Bia lystok, Poland
101
SUPA, University of Strathclyde, Glasgow G1 1XQ, United Kin
gdom
102
IISER-TVM, CET Campus, Trivandrum Kerala 695016, India
103
Northwestern University, Evanston, IL 60208, USA
104
Institute of Applied Physics, Nizhny Novgorod, 603950, Rus
sia
105
Pusan National University, Busan 609-735, Korea
106
Hanyang University, Seoul 133-791, Korea
107
NCBJ, 05-400
́
Swierk-Otwock, Poland
108
IM-PAN, 00-956 Warsaw, Poland
109
Rochester Institute of Technology, Rochester, NY 14623, US
A
110
Monash University, Victoria 3800, Australia
111
Seoul National University, Seoul 151-742, Korea
112
ESPCI, CNRS, F-75005 Paris, France
113
Universit`a di Camerino, Dipartimento di Fisica, I-62032 C
amerino, Italy
114
Southern University and A&M College, Baton Rouge, LA 70813,
USA
115
College of William and Mary, Williamsburg, VA 23187, USA
116
Instituto de F ́ısica Te ́orica, University Estadual Paulis
ta/ICTP South
American Institute for Fundamental Research, S ̃ao Paulo SP
01140-070, Brazil
117
IISER-Kolkata, Mohanpur, West Bengal 741252, India
118
Rutherford Appleton Laboratory, HSIC, Chilton, Didcot, Ox
on OX11 0QX, United Kingdom
119
Whitman College, 280 Boyer Ave, Walla Walla, WA 9936, USA
120
National Institute for Mathematical Sciences, Daejeon 305
-390, Korea
121
Hobart and William Smith Colleges, Geneva, NY 14456, USA
122
Institute of Astronomy, 65-265 Zielona G ́ora, Poland
123
Andrews University, Berrien Springs, MI 49104, USA
124
Universit`a di Siena, I-53100 Siena, Italy
125
Trinity University, San Antonio, TX 78212, USA
126
University of Washington, Seattle, WA 98195, USA and
127
Abilene Christian University, Abilene, TX 79699, USA
(Dated: 12th February 2016)
We present the results of a search for long-duration gravita
tional wave transients in two sets of
data collected by the LIGO Hanford and LIGO Livingston detec
tors between November 5, 2005
and September 30, 2007, and July 7, 2009 and October 20, 2010,
with a total observational time of
283.0 days and 132.9 days, respectively. The search targets
gravitational wave transients of duration
10–500 s in a frequency band of 40–1000 Hz, with minimal assum
ptions about the signal waveform,
polarization, source direction, or time of occurrence. All
candidate triggers were consistent with the
expected background; as a result we set 90% confidence upper l
imits on the rate of long-duration
gravitational wave transients for different types of gravit
ational wave signals. For signals from black
hole accretion disk instabilities, we set upper limits on th
e source rate density between 3
.
4
×
10
5
9
.
4
×
10
4
Mpc
3
yr
1
at 90% confidence. These are the first results from an all-sky s
earch for
unmodeled long-duration transient gravitational waves.
I. INTRODUCTION
The goal of the Laser Interferometer Gravitational-
Wave Observatory (LIGO) [1] and the Virgo detect-
ors [2] is to directly detect and study gravitational waves
(GWs). The direct detection of GWs holds the prom-
publication@ligo.org; publication@ego-gw.it
ise of testing general relativity in the strong-field regime,
of providing a new probe of objects such as black holes
and neutron stars, and of uncovering unanticipated new
astrophysics.
LIGO and Virgo have jointly acquired data that have
been used to search for many types of GW signals: un-
modeled bursts of short duration (
<
1 s) [3–7], well-
modeled chirps emitted by binary systems of compact
objects [8–12], continuous signals emitted by asymmet-
ric neutron stars [13–20], as well as a stochastic back-
6
ground of GWs [21–24]. For a complete review, see [25].
While no GW sources have been observed by the first-
generation network of detectors, first detections are ex-
pected with the next generation of ground-based detect-
ors: advanced LIGO [26], advanced Virgo [27], and the
cryogenic detector KAGRA [28]. It is expected that the
advanced detectors, operating at design sensitivity, will
be capable of detecting approximately 40 neutron star
binary coalescences per year, although significant uncer-
tainties exist [29].
Previous searches for unmodeled bursts of GWs [3–
5] targeted source objects such as core-collapse super-
novae [30], neutron star to black hole collapse [31], cos-
mic string cusps [32], binary black hole mergers [33–35],
star-quakes in magnetars [36], pulsar glitches [37], and
signals associated with gamma ray bursts (GRBs) [38].
These burst searches typically look for signals of duration
1 s or shorter.
At the other end of the spectrum, searches for persist-
ent, unmodeled (stochastic) GW backgrounds have also
been conducted, including isotropic [21], anisotropic and
point-source backgrounds [22]. This leaves the parameter
space of unmodeled transient GWs not fully explored; in-
deed, multiple proposed astrophysical scenarios predict
long-duration GW transients lasting from a few seconds
to hundreds of seconds, or even longer, as described in
Section II. The first search for unmodeled long-duration
GW transients was conducted using LIGO data from the
S5 science run, in association with long GRBs [39]. In
this paper, we apply a similar technique [40] in order to
search for long-lasting transient GW signals over all sky
directions and for all times. We utilize LIGO data from
the LIGO Hanford and Livingston detectors from the S5
and S6 science runs, lasting from November 5, 2005 to
September 30, 2007 and July 7, 2009 to October 20, 2010,
respectively.
The organization of the paper is as follows. In Sec-
tion II, we summarize different types of long-duration
transient signals which may be observable by LIGO and
Virgo. In Section III, we describe the selection of the
LIGO S5 and S6 science run data that have been used
for this study. We discuss the search algorithm, back-
ground estimation, and data quality methods in Section
IV. In Section V, we evaluate the sensitivity of the search
to simulated GW waveforms. The results of the search
are presented in Section VI. We conclude with possible
improvements for a long-transient GW search using data
from the advanced LIGO and Virgo detectors in Section
VII.
II. ASTROPHYSICAL SOURCES OF LONG GW
TRANSIENTS
Some of the most compelling astrophysical sources of
long GW transients are associated with extremely com-
plex dynamics and hydrodynamic instabilities following
the collapse of a massive star’s core in the context of core-
collapse supernovae and long GRBs [30, 40, 41]. Soon
after core collapse and the formation of a proto-neutron
star, convective and other fluid instabilities (including
standing accretion shock instability [42]) may develop be-
hind the supernova shock wave as it transitions into an
accretion shock. In progenitor stars with rapidly rotat-
ing cores, long-lasting, non-axisymmetric rotational in-
stabilities can be triggered by corotation modes [43–46].
Long-duration GW signals are expected from these viol-
ently aspherical dynamics, following within tens of milli-
seconds of the short-duration GW burst signal from core
bounce and proto-neutron star formation. Given the tur-
bulent and chaotic nature of post-bounce fluid dynamics,
one expects a stochastic GW signal that could last from
a fraction of a second to multiple seconds, and possibly
even longer [30, 40, 47–50].
After the launch of an at least initially successful ex-
plosion, fallback accretion onto the newborn neutron star
may spin it up, leading to non-axisymmetric deforma-
tion and a characteristic upward chirp signal (700 Hz–
few kHz) as the spin frequency of the neutron star in-
creases over tens to hundreds of seconds [51, 52]. GW
emission may eventually terminate when the neutron star
collapses to a black hole. The collapse process and form-
ation of the black hole itself will also produce a short-
duration GW burst [31, 53, 54].
In the collapsar model for long GRBs [55], a stellar-
mass black hole forms, surrounded by a massive, self-
gravitating accretion disk. This disk may be sus-
ceptible to various non-axisymmetric hydrodynamic and
magneto-hydrodynamic instabilities which may lead to
fragmentation and inspiral of fragments into the central
black hole (e.g., [56, 57]). In an extreme scenario of such
accretion disk instabilities (ADIs), magnetically “suspen-
ded accretion” is thought to extract spin energy from the
black hole and dissipate it via GW emission from non-
axisymmetric disk modes and fragments [58, 59]. The as-
sociated GW signal is potentially long-lasting (10–100s)
and predicted to exhibit a characteristic downward chirp.
Finally, in magnetar models for long and short GRBs
(e.g., [60, 61]), a long-lasting post-GRB GW transient
may be emitted by a magnetar undergoing rotational or
magnetic non-axisymmetric deformation (e.g., [62, 63]).
III. DATA SELECTION
During the fifth LIGO science run (S5, November 5,
2005 to September 30, 2007), the 4 km and 2 km detect-
ors at Hanford, Washington (H1 and H2), and the 4 km
detector at Livingston, Louisiana (L1), recorded data for
nearly two years. They were joined on May 18, 2007
by the Virgo detector (V1) in Pisa, Italy, which was be-
ginning its first science run. After a two-year period of
upgrades to the detectors and the decommissioning of
H2, the sixth LIGO and second and third Virgo scientific
runs were organized jointly from July 7, 2009 to October
10, 2010.
7
Among the four detectors, H1 and L1 achieved the
best strain sensitivity, reaching
2
×
10
23
/
Hz around
150 Hz in 2010 [64, 65]. Because of its reduced arm
length, H2 sensitivity was at least a factor of 2 lower
than H1 on average. V1 sensitivity varied over time, but
was always lower than the sensitivity of H1 and L1 by
a factor between 1.5 and 5 at frequencies higher than
60 Hz. Moreover, the H1-L1 pair livetime was at least a
factor 2 longer than the livetime of the H1-V1 and L1-V1
pairs added together. Using Virgo data, however, could
help with sky localization of source candidates; unfortu-
nately, the sky localization was not implemented at the
time of this search. Consequently, including Virgo data
in this analysis would have increased the overall search
sensitivity by only a few percent or less at the cost of
analyzing two additional pairs of detectors. As a result,
we have analyzed only S5 and S6 data from the H1-L1
pair for this search.
In terms of frequency content, we restrict the analysis
to the 40–1000Hz band. The lower limit is constrained
by seismic noise, which steeply increases at lower frequen-
cies in LIGO data. The upper limit is set to include the
most likely regions of frequency space for long-transient
GWs, while keeping the computational requirements of
the search at an acceptable level. We note that the fre-
quency range of our analysis includes the most sensitive
band of the LIGO detectors, namely 100–200Hz.
Occasionally, the detectors are affected by instru-
mental problems (data acquisition failures, misalignment
of optical cavities, etc.) or environmental conditions (bad
weather, seismic motion, etc.) that decrease their sensit-
ivity and increase the rate of data artifacts, or glitches.
Most of these periods have been identified and can be
discarded from the analysis using data quality flags [66–
69]. These are classified by each search into different
categories depending on how the GW search is affected.
Category 1 data quality flags are used to define peri-
ods when the data should not be searched for GW sig-
nals because of serious problems, like invalid calibration.
To search for GW signals, the interferometers should be
locked and there should be no evidence of environmental
noise transients corrupting the measured signal. For this
search, we have used the category 1 data quality flags
used by searches for an isotropic stochastic background
of GWs [21, 23]. This list of flags is almost identical to
what has been used by the unmodeled all-sky searches
for short-duration GW transients [3, 4]. We also discard
times when simulated signals are injected into the detect-
ors through the application of a differential force onto the
mirrors.
Category 2 data quality flags are used to discard trig-
gers which pass all selection cuts in a search, but are
clearly associated with a detector malfunction or an en-
vironmental condition [68]. In Section IV C, we explain
which category 2 flags have been selected and how we use
them in this search.
Overall, we discard 5.8% and 2.2% of H1-L1 coincident
data with our choices of category 1 data quality flags
for S5 and S6, respectively. The remaining coincident
strain time series are divided into 500 s intervals with 50%
overlap. Intervals smaller than 500 s are not considered.
For the H1-L1 pair, this results in a total observation
time of 283.0 days during S5 and 132.9 daysfor S6.
IV. LONG TRANSIENT GW SEARCH
PIPELINE
A. Search algorithm
The search algorithm we employ is based on the cross-
correlation of data from two GW detectors, as described
in [40]. This algorithm builds a frequency-time map (
ft
-
map) of the cross-power computed from the strain time-
series of two spatially separated detectors. A pattern
recognition algorithm is then used to identify clusters
of above-threshold pixels in the map, thereby defin-
ing candidate triggers. A similar algorithm has been
used to search for long-lasting GW signals in coincid-
ence with long GRBs in LIGO data [39]. Here we ex-
tend the method to carry out an un-triggered (all-sky,
all-time) search, considerably increasing the parameter
space covered by previous searches.
Following [40], each 500 s interval of coincident data is
divided into 50% overlapping, Hann-windowed, 1 s long
segments. Strain data from each detector in the given
1 s segment are then Fourier transformed, allowing form-
ation of
ft
-maps with a pixel size of 1 s
×
1 Hz. An
estimator for GW power can be formed [40]:
ˆ
Y
(
t
;
f
;
ˆ
Ω) =
2
N
Re
h
Q
IJ
(
t
;
f
;
ˆ
Ω) ̃
s
I
(
t
;
f
) ̃
s
J
(
t
;
f
)
i
.
(1)
Here
t
is the start time of the pixel,
f
is the frequency
of the pixel,
ˆ
Ω is the sky direction,
N
is a window nor-
malization factor, and ̃
s
I
and ̃
s
J
are the discrete Fourier
transforms of the strain data from GW detectors
I
and
J
. We use the LIGO H1 and L1 detectors as the
I
and
J
detectors, respectively. The optimal filter
Q
IJ
takes into
account the phase delay due to the spatial separation
of the two detectors, ∆
~x
IJ
, and the direction-dependent
efficiency of the detector pair,
ǫ
IJ
(
t
;
ˆ
Ω):
Q
IJ
(
t
;
f
;
ˆ
Ω) =
e
2
πif
~x
IJ
·
ˆ
/c
ǫ
IJ
(
t
;
ˆ
Ω)
.
(2)
The pair efficiency is defined by:
ǫ
IJ
(
t
;
ˆ
Ω) =
1
2
X
A
F
A
I
(
t
;
ˆ
Ω)
F
A
J
(
t
;
ˆ
Ω)
,
(3)
where
F
A
I
(
t
;
ˆ
Ω) is the antenna factor for detector
I
and
A
is the polarization state of the incoming GW [40]. An
estimator for the variance of the
ˆ
Y
(
t
;
f,
ˆ
Ω) statistic is
then given by:
ˆ
σ
2
Y
(
t
;
f
;
ˆ
Ω) =
1
2
|
Q
IJ
(
t
;
f
;
ˆ
Ω)
|
2
P
adj
I
(
t
;
f
)
P
adj
J
(
t
;
f
)
,
(4)
8
where
P
adj
I
(
t
;
f
) is the average one-sided power spectrum
for detector
I
, calculated by using the data in 8 non-
overlapping segments on each side of time segment
t
[40].
We can then define the cross-correlation signal-to-noise
ratio (SNR) in a single pixel,
ρ
:
ρ
(
t
;
f
;
ˆ
Ω) =
ˆ
Y
(
t
;
f
;
ˆ
Ω)
/
ˆ
σ
Y
(
t
;
f
;
ˆ
Ω)
.
(5)
Because this is proportional to strain squared, it is an
energy SNR, rather than an amplitude SNR.
This statistic is designed such that true GW signals
should induce positive definite
ρ
when the correct filter
is used (i.e., the sky direction
ˆ
Ω is known). Consequently,
using a wrong sky direction in the filter results in reduced
or even negative
ρ
for real signals. Figure 1 shows an
example
ft
-map of
ρ
containing a simulated GW signal
with a known sky position.
Time [s]
Frequency [Hz]
20
40
60
80
100
150
200
250
ρ
−5
0
5
Figure 1:
ft
-map of
ρ
(cross-correlation signal-to-noise ratio)
using simulated Gaussian data. A simulated GW signal from
an accretion disk instability [58, 59] (model waveform ADI-
E, see Table I) with known sky position is added to the data
stream and is visible as a bright, narrow-band track. Blurri
ng
around the track is due to the usage of adjacent time segments
in estimating ˆ
σ
Y
; the estimate of ˆ
σ
Y
in these bins is affected
by the presence of the GW signal.
Next, a seed-based clustering algorithm [70] is applied
to the
ρ ft
-map to identify significant clusters of pixels.
In particular, the clustering algorithm applies a threshold
of
|
ρ
| ≥
1 to identify seed pixels, and then groups these
seed pixels that are located within a fixed distance (two
pixels) of each other into a cluster. These parameters
were determined through empirical testing with simu-
lated long-transient GW signals similar to those used in
this search (discussed further in Section V A). The res-
ulting clusters (denoted Γ) are ranked using a weighted
sum of the individual pixel values of
ˆ
Y
and ˆ
σ
Y
:
SNR
Γ
(
ˆ
Ω) =
X
t
;
f
Γ
ˆ
Y
(
t
;
f
;
ˆ
Ω)ˆ
σ
2
Y
(
t
;
f
;
ˆ
Ω)
X
t
;
f
Γ
ˆ
σ
2
Y
(
t
;
f
;
ˆ
Ω)
1
/
2
.
(6)
SNR
Γ
(
ˆ
Ω) represents the signal-to-noise ratio of the
cluster Γ.
In principle, this pattern recognition algorithm could
be applied for every sky direction
ˆ
Ω, since each sky di-
rection is associated with a different filter
Q
IJ
(
t
;
f
;
ˆ
Ω).
However, this procedure is prohibitively expensive from
a computational standpoint. We have therefore modified
the seed-based clustering algorithm to cluster both pixels
with positive
ρ
and those with negative
ρ
(arising when
an incorrect sky direction is used in the filter). Since
the sky direction is not known in an all-sky search, this
modification allows for the recovery of some of the power
that would normally be lost due to a suboptimal choice
of sky direction in the filter.
The algorithm is applied to each
ft
-map a certain num-
ber of times, each iteration corresponding to a different
sky direction. The sky directions are chosen randomly,
but are fixed for each stretch of uninterrupted science
data. Different methods for choosing the sky directions
were studied, including using only sky directions where
the detector network had high sensitivity and choosing
the set of sky directions to span the set of possible sig-
nal time delays. The results indicated that sky direction
choice did not have a significant impact on the sensitivity
of the search.
We also studied the effect that the number of sky dir-
ections used had on the search sensitivity. We found
that the search sensitivity increased approximately log-
arithmically with the number of sky directions, while the
computational time increased linearly with the number
of sky directions. The results of our empirical studies
indicated that using five sky directions gave the optimal
balance between computational time and search sensit-
ivity.
This clustering strategy results in a loss of sensitivity
of
10–20% for the waveforms considered in this search
as compared to a strategy using hundreds of sky direc-
tions and clustering only positive pixels. However, this
strategy increases the computational speed of the search
by a factor of 100 and is necessary to make the search
computationally feasible.
We also apply two data cleaning techniques concur-
rently with the data processing. First, we remove fre-
quency bins that are known to be contaminated by in-
strumental and environmental effects. This includes the
violin resonance modes of the suspensions, power line
harmonics, and sinusoidal signals injected for calibration
purposes. In total, we removed 47 1 Hz-wide frequency
bins from the S5 data, and 64 1 Hz-wide frequency bins
from the S6 data. Second, we require the waveforms ob-
9
served by the two detectors to be consistent with each
other, so as to suppress instrumental artifacts (glitches)
that affect only one of the detectors. This is achieved by
the use of a consistency-check algorithm [71] which com-
pares the power spectra from each detector, taking into
account the antenna factors.
B. Background estimation
An important aspect of any GW search is understand-
ing the background of accidental triggers due to detector
noise; this is crucial for preventing false identification of
noise triggers as GW candidates. To estimate the false
alarm rate (FAR), i.e. the rate of accidental triggers due
to detector noise, we introduce a non-physical time-shift
between the H1 and L1 strain data before computing
ρ
.
Analysis of the time-shifted data proceeds identically to
that of unshifted data (see Section IV A for more details).
With this technique, and assuming the number of hypo-
thetical GW signals is small, the data should not contain
a correlated GW signal, so any triggers will be generated
by the detector noise. We repeat this process for multiple
time-shifts in order to gain a more accurate estimate of
the FAR from detector noise.
As described in Section III, each analysis segment is
divided into 500 s long intervals which overlap by 50%
and span the entire dataset. For a given time-shift
i
,
the H1 data from interval
n
are correlated with L1 data
from the interval
n
+
i
. Since the time gap between two
consecutive intervals may be non-zero, the actual time-
shift applied in this process is at least 500
×
i
seconds. The
time-shift is also circular: if for a time-shift
i
,
n
+
i > N
(where
N
is the number of overlapping intervals required
to span the dataset), then H1 data from the interval
n
are
correlated with L1 data from the interval
n
+
i
N
. It is
important to note that the minimum time-shift duration
is much longer than the light travel time between the
two detectors and also longer than the signal models we
consider (see Section V for more information) in order to
prevent accidental correlations.
Using this method, 100 time-shifts have been processed
to estimate the background during S5, amounting to a
total analyzed livetime of 84.1 years. We have also stud-
ied 100 time-shifts of S6 data, with a total analyzed liv-
etime of 38.7 years. The cumulative rates of background
triggers for the S5 and S6 datasets can be seen in Fig-
ures 2 and 3, respectively.
C. Rejection of loud noise triggers
As shown in Figures 2 and 3, the background FAR dis-
tribution has a long tail extending to SNR
Γ
>
100; this
implies that detector noise alone can generate triggers
containing significant power. Many of these triggers are
caused by short bursts of non-stationary noise (glitches)
in H1 and/or L1, which randomly coincide during the
time-shifting procedure. It is important to suppress these
types of triggers so as to improve the significance of true
GW signals in the unshifted data.
These glitches are typically much less than 1 s in dura-
tion, and as a result, nearly all of their power is concen-
trated in a single 1 s segment. To suppress these glitches,
we have defined a discriminant variable, SNRfrac, that
measures the fraction of SNR
Γ
located in a single time
segment. The same SNRfrac threshold of 0.45 was found
to be optimal for all simulated GW waveforms using both
S5 and S6 data. This threshold was determined by max-
imizing the search sensitivity for a set of simulated GW
signals (see Section V); we note that this was done before
examining the unshifted data, using only time-shifted
triggers and simulated GW signals. The detection effi-
ciency is minimally affected (less than 1%) by this SNR-
frac threshold choice.
We also utilize LIGO data quality flags to veto triggers
generated by a clearly identified source of noise. We have
considered all category 2 data quality flags used in un-
modeled or modeled transient GW searches [68]. These
flags were defined using a variety of environmental mon-
itors (microphones, seismometers, magnetometers, etc.)
and interferometer control signals to identify stretches of
data which may be compromised due to local environ-
mental effects or instrumental malfunction. Since many
of these data quality flags are not useful for rejecting
noise triggers in this analysis, we select a set of effect-
ive data quality flags by estimating the statistical sig-
nificance of the coincidence between these data quality
flags and the 100 loudest triggers from the time-shifted
background study (no SNRfrac selection applied). The
significance is defined by comparing the number of coin-
cident triggers with the accidental coincidence mean and
standard deviation. Given the small number of triggers
we are considering (100), and in order to avoid accidental
coincidence, we have applied a stringent selection: only
those data quality flags which have a statistical signi-
ficance higher than 12 standard deviations (as defined
above) and an efficiency over deadtime ratio larger than 8
have been selected. Here, efficiency refers to the fraction
of noise triggers flagged, while deadtime is the amount of
science data excluded by the flag.
For both the S5 and S6 datasets, this procedure se-
lected data quality flags which relate to malfunctions of
the longitudinal control of the Fabry-Perot cavities and
those which indicate an increase in seismic noise. The
total deadtime which results from applying these data
quality (DQ) flags amounts to
12 hours in H1 and L1
(0
.
18%) for S5 and
4 hours in H1 (0
.
13%) and
7 hours
in L1 (0
.
22%) for S6.
As shown in Figures 2 and 3, these two data quality
cuts (SNRfrac and DQ flags) are useful for suppressing
the high-SNR
Γ
tail of the FAR distribution. More pre-
cisely, the SNRfrac cut is very effective for cleaning up
coincident glitches with high SNR
Γ
, while the DQ flags
are capable of removing less extreme triggers occurring
due to the presence of a well-identified noise source. We