The LIGO Scientific Collaboration and the Virgo Collaboration report:
We performed a preliminary offline analysis using the PyCBC search (Nitz
et al. arxiv:1705.01513, 2017) of the binary neutron star candidate
G298048 (LSC and Virgo, GCN 21505, 21509, 21510) identified in low-latency
by the gstlal online search (Messick et al. Phys. Rev. D 95, 042001, 2017).
A trigger consistent with a binary neutron star merger is observed at GPS
time 1187008882.443 (2017-08-17 12:41:04 UTC) in both the LIGO Livingston
(L1) and LIGO Hanford (H1) detectors. The trigger is below threshold in
Virgo because of the antenna pattern for Virgo (V1) at the time and
location of this event, but the Virgo instrument contributes to the
localization. The duration of the gravitational-wave signal is
approximately 74 seconds from the search���s low-frequency cutoff of 27 Hz
to the binary merger.
Investigation of L1 data identified a noise transient from a known class
of instrumental glitches during the inspiral signal. The duration of this
glitch is a small fraction of a second and does not appear to affect the
signal at times away from the glitch. To make an improved preliminary
estimate of the sky position, we re-analyzed the data, removing the L1
noise transient at GPS time 1187008881.389 by multiplying the strain data
with a Tukey window, such that the total duration of the zeroed data is
0.2 s and the total duration of the Tukey window is 1.2 s.
An updated BAYESTAR sky map (Singer et al. 2016, ApJL 829, 15) that uses
data from all three gravitational-wave observatories (H1, L1, and V1) is
available for retrieval from the GraceDB page
(https://gracedb.ligo.org/events/view/G298048): bayestar-HLV.fits.gz. The
centroid (maximum a posteriori) sky location is R.A.=12h57m, Dec.=-17d51m.
The 50% credible region spans about 9 deg2 and the 90% region about 31
deg2. The luminosity distance is 40 +/- 8 Mpc (all-sky a posteriori mean
+/- standard deviation). This is the preferred sky map at this time.
If we assume that the binary is either face on or face off in the plane of
the sky, then we obtain a revised estimate of luminosity distance of 50
+/- 3 Mpc. This assumption does not significantly affect the overall 2D
localization projected onto the sky, but reduces the 3D volume of the
localization (e.g. Pankow et al. 2017, ApJ 834, 154).
We caution that the parameters and significance of this candidate may be
subject to change as data-quality, calibration, and full parameter
estimation studies are ongoing.