How many short duration Gamma Ray Bursts could be giant flares from Soft Gamma Repeaters R Chapman

1
How many short duration Gamma Ray Bursts could
be giant flares fromSoft Gamma Repeaters?R
Chapman N R Tanvir, University of
Hertfordshire(email bchapman_at_star.herts.ac.uk)
Short Bursts and Giant Flares Gamma Ray Bursts
with durations less than 2s form a distinct
population1 whose origin remains a mystery. On
27th December 2004, a giant flare from the Soft
Gamma Repeater SGR1806-20 saturated all
spaceborne gamma ray detectors2. Particle
detectors aboard the RHESSI spacecraft provided a
lower limit on its fluence suggesting a total
energy release of 0.8 x 1046 erg at a distance of
15 kpc. As is characteristic for SGR giant
flares, this flare began with a short (0.2s),
intense spike, followed by a long (380s)
pulsating tail. It has been suggested that the
initial spike closely resembles the temporal
behaviour of short GRBs, and given the energy
estimates above, giant flares could have been
detected by BATSE out to distances in the range
30-80Mpc2,3,4,5. Here we analyse correlations of
short-duration bursts in the BATSE 4B catalogue
with the IRAS PSCz catalogue6 and the Third
Reference Catalogue of Bright Galaxies7. Two
point angular correlations and a weighted
statistic for overall correlation are computed
for various duration GRBs versus sets of galaxies
at three limiting recession velocities (1500,
2000 and 5000 kms-1 corresponding to 21, 28 and
70 Mpc (assuming H071kms-1Mpc-1).
The Real Datasets The BATSE 4B catalogue contains
2039 Gamma Ray Bursts of known duration (T90
measuring the time interval during which 90 of
the total counts for each burst were received).
In order to investigate any correlation of short
bursts with relatively local galactic
structure, we selected short bursts of varying
T90 durations (0.2, 0.5, 1.0 and 2.0 s with T90
errors lt 50 and positional error circles of
radius lt 8 degrees). Correlation analyses were
then performed against sets of galaxies with
recession velocities lt 1500, 2000 and 5000 kms-1
from the IRAS PSCz and RC3 catalogues.
Pseudo-burst Datasets In order to examine the
level of correlation which could be detected by
this method, test datasets of pseudo-bursts were
generated by selecting random galaxies from each
catalogue, smearing each of their positions by a
randomly chosen BATSE error circle, and building
test datasets with 0-100 correlation. Example
correlation plots from these are shown in Figure
1 and 2.
Angular Correlation Two point correlation
functions were computed for each dataset, where
the angular distance of every burst from every
galaxy was measured, placed into 4 degree bins,
counted and compared with a distribution of
10,000 random galaxies (e.g.Figure 3).
BigSum Correlations Comparing Figures 1,2 and 3
it can be seen that angular correlation plots
imply that the fraction of correlated short
bursts is less than 20. To further quantify this
upper limit, we define a total BigSum
correlation as shown in equation 1. For every
burst, the distance (weighted by its Gaussian
error circle) to every galaxy is calculated, and
an overall sum for the entire set is calculated
and compared with the sum for the 10,000 random
galaxy distribution. Plots of BigSum for 0-100
pseudo-burst datasets (of equal numbers of
pseudo-burst to real burst) were created, and the
line of best fit calculated with 2-sigma (95)
confidence limits (e.g. Figure 4). From these
calibration plots, an upper limit of the fraction
of real bursts correlated with each galaxy
distribution can be determined by measuring the
BigSum for real burst datasets. These results are
plotted as Figures 5 and 6.

• References
• Kouveliotou, C., et al. Astrophys. J, 413,
  L101-L104 (1993)
• Hurley, K., et al. Nature 2005 in press
  (astro-ph/0502329)
• Dar, A., 2005, GCN circ.2942
• Nakar et al., astro-ph/0502148
• Popov, S.B., astro-ph/0502391
• Saunders, W., et al. MNRAS, 317, 55 (2000)
• de Vaucouleurs et al, Springer-Verlag New York
  (1991)

Figure 5
Figure 6
Conclusions We have analysed the entire BATSE
catalogue and compared short burst distributions
with varying sized shells of local galaxies. With
these results we can provide 95 confident
upper limits on the fraction of BATSE short
bursts associated with local extra-galactic
structure, and therefore possibly giant
flares from SGRs, as a function of both burst
duration and galaxy distance. For example, out to
a distance of 28 Mpc (recession velocity 2000
kms-1 in PSCz)
Burst Upper limit on fraction of bursts
due T90 to Soft Gamma Repeater giant
flares ? 0.2s 23 ? 0.5s 15 ?
1.0s 18 ? 2.0s 20