Transcript XLII RENCONTRES DE MORIOND WORKSHOP ON …

44th Rencontres de Moriond
Very High Energy Phenomena in the Universe
Why the Swift GRB redshift distribution is
changing in time
Dr David Coward & Alan Imerito
University of Western Australia
01/02/2009
Moriond 2009
The GRB redshift distribution
• The spatial distribution of
GRBs is a powerful probe of
GRB rate evolution.
• Potentially be used as an
independent tracer of massive
star formation in the early
Universe.
• Potentially be used to probe
the evolution of GRB
environments.
01/02/2009
Moriond 2009
What we observe
• Optical/NIR afterglows have been found for nearly
80% of GRBs
• Only 40–50% of these have measured redshifts (now
over 100)
• Optically dark bursts - extinctionGRB environment, host galaxy type and distance
• Preferentially measure redshifts from optically bright
GRBs.
01/02/2009
Moriond 2009
GRB redshift statistics
• Pre-Swift - <z> about 1.4
• Early Swift - <z> about 2.8 (in 2005-2006)
• Swift is more sensitive to higher-z longer duration
GRBs
• Recent Swift -<z> about 2 (2008)
• Statistical moments of the redshift distribution should
converge to constants given enough statistics
01/02/2009
Moriond 2009
Time series analysis of GRB redshifts
• Search for evolution in statistical moments
• time-dependent selection effects
• 92 long GRB redshifts from 2005-2008
• Motivation - redshifts measured from bright
optical afterglow absorption spectra - expect
biases
• This is linked to the efficiency of GRB follow-up
telescopes to acquire absorption spectra
01/02/2009
Moriond 2009
Swift triggered redshift time series
Red squares - redshifts
Solid line - nearest
neighbour averaging of
reshift
Open circles - optical
afterglow magnitude at
discovery
Average z is evolving
on time-scale of years must be an observation
bias
01/02/2009
Moriond 2009
Response times for spectroscopy
enabled GRB follow-up telescopes
Optical afterglow
brightness decays as 1/T
Average time to acquire
absorption spectra for the
VLT has reduced from
about 1000 min in 2005
to 100 min in 2008
The so-called learning
curve effect
01/02/2009
Moriond 2009
Raw correlation between
telescope response time
and redshift
Pearson Rho = 0.41
Plotting the time-series
windowed averages clearly
identifies the +ve trend
More probable for a long
response time to be
correlated with a more
distant GRB.
Does this make sense!
01/02/2009
Moriond 2009
Malmquist bias revisited
• Malmquist bias - for flux limited surveys - high-z events
originate from the bright end of the GRB optical LF.
• At small-z, can see both faint and bright end of LF.
• Long telescope response times -> fainter OA because
of 1/T -> only seen at relatively smaller z
• Short telescope response times -> brighter OA -> seen
at relatively higher z
• What we find is the opposite…an “anti-Malmquist” bias!
01/02/2009
Moriond 2009
Non-evolving simulated GRB optical LF to
demonstrate the Malmquist bias
01/02/2009
Moriond 2009
Simulated Malmquist bias on average redshift for
different telescope response times
Malmquist
Anti-Malmquist
Long response
times (fainter OA)
correlated with
smaller redshifts
01/02/2009
Moriond 2009
Response times plotted with average redshift of the
potentially observable OA…using an evolving OA LF.
Anti-Malmquist
01/02/2009
Moriond 2009
To produce an anti-Malmquist bias in the simulations
we employ an OA LF that evolves with z.
GRBs OA optical
brightness must be
evolving with z?
Are the high-z
bursts intrinsically
brighter or less
obscured?
01/02/2009
Moriond 2009
Summary
• Response times of large telescopes to acquire a redshift are
decreasing in 2005-2008 period
• Average GRB redshift is reducing over the same period
• Longer average telescope response times are correlated with
larger average redshifts
• An “anti-Malmquist” bias is observed: that is GRBs at high-z
are easier to see than expected
• To reconcile this trend, simulations suggest that GRBs at high-z
must be relatively brighter than those at small-z
• The analysis implies that GRB optical selection effects are
potentially an important tool for probing GRB environments
01/02/2009
Moriond 2009
Future work
• Use OA data to confirm how the OA brightness affects the
probability of obtaining a redshift
• Differentiate between dust obscuration and intrinsic GRB
brightness
• Is the change of GRB optical obscuration with z linked
with the history of massive star evolution?
• Selection effects in astronomy are often considered a
problem…in this case they might actually reveal new
insight into the origin and evolution of of GRBs
01/02/2009
Moriond 2009
Acknowledgements
• Australian Research Council
• UWA
• Moriond 09 organisers - Hady schenten and
many others
• I have only seen snow twice in my
lifetime…Moriond 09 is the second time
01/02/2009
Moriond 2009