TITLE:   GCN CIRCULAR
NUMBER:  11854
SUBJECT: GRB 110328A / Swift J164449.3+573451: Radio-optical/NIR Astrometry
DATE:    11/03/31 18:47:00 GMT
FROM:    Edo Berger at Harvard  <eberger@cfa.harvard.edu>

E. Berger (Harvard), A. Levan (U. Warwick), N. R. Tanvir
(U. Leicester), A. Zauderer, A. M. Soderberg (Harvard), and
D. A. Frail (NRAO) report:

"We performed absolute and relative astrometry on radio and optical
images of GRB 110328A / Swift J164449.3+573451 (GCNs 11823, 11824) in
an atempt to locate the radio transient (GCNs 11836, 11848) relative
to the quiescent optical source at z=0.35 (GCNs 11827, 11830, 11833,
11834).  We used EVLA observations at a mean frequency of 5.8 GHz
(GCNs 11836, 11848), a Gemini-N/GMOS r-band image, and a UKIRT/WFCAM
J-band image (GCN 11846).

The Gemini and UKIRT images were astrometrically aligned relative to
2MASS using 20 and 55 objects in common, respectively, leading to an
rms scatter of 0.14" and 0.10" in each coordinate, respectively.  For
the EVLA image we used the native astrometric solution relative to the
ICRS.  The accuracy of the astrometry is verified by imaging and
measuring the position of the phase calibrator, which agrees with the
ICRS position to better than 0.4 mas.  We note that the 2MASS
point-source-catalog positions used to align the Gemini and UKIRT
images are reconstructed onto the ICRS.

We find the following coordinates for the optical source in the Gemini
image (J2000):
      RA = 16:44:49.939
      DEC = +57:34:59.64
with an uncertainty of 0.14" in each coordinate, dominated by the
astrometric solution uncertainty.

The position of the faint NIR counterpart in the UKIRT image is:
      RA = 16:44:49.958
      DEC = +57:35:00.00
with an uncertainty of about 0.3" in the source centroid, and an
additional 0.1" due to astrometric solution uncertainty.

Finally, the ICRS position of the radio transient is:
      RA = 16:44:49.925
      DEC = +57:34:59.68
with negligible centroid uncertainty of about 0.01".

The relative offset between the radio and optical positions is
therefore 0.12+/-0.20", or about 0.6+/-1.0 kpc (at z=0.35).  The
offset relative to the NIR position is 0.42+/-0.42".  Thus, the radio
transient position is consistent with arising in the nucleus of the
host galaxy (GCN 11847).

We find an additional common source between the radio, optical, and
NIR images with the following positions:

optical:
      RA = 16:44:48.047
      DEC = +57:32:16.83
near-IR:
      RA = 16:44:48.093
      DEC = +57:32:17.03
radio:
      RA = 16:44:48.139
      DEC = +57:32:16.74

This source is a galaxy with extended structure in the optical/NIR
images, but it is unresolved in the radio.  We therefore associate the
radio position with the nucleus of the galaxy (positions given above).
The resulting positional offsets are 0.75+/-0.20" (radio-to-optical)
and 0.47+/-0.15" (radio-to-NIR).  We caution that the error bars do
not include a systematic uncertainty in centroiding on the galaxy
nucleus (~0.1"), and do not account for potential distortions due to
the location of the object near the edge of the Gemini image.  Taken
at face value, the offset between these positions potentially reflects
a relative shift between the optical/NIR and radio astrometric
systems.  If applied to the radio counterpart of GRB 110328A / Swift
J164449.3+573451, this shift would lead to a significant offset
(~3-sigma confidence level) between the radio position and the center
of the host galaxy.

To conclude, absolute astrometry indicates that the radio transient is
consistent with arising in the nucleus of the host galaxy (0.6+/-1.0
kpc); the relative radio-optical/NIR positions of an additional common
source allow for an offset of up to ~3 kpc.  Additional wide-field
(~10') deep NIR imaging may provide a larger number of common objects
with the EVLA images, thereby impoving the relative astrometric
solution.