J/ApJ/748/134 Variability components in BATSE GRB light curves (Gao+, 2012)
Stepwise filter correlation method and evidence of superposed variability
components in gamma-ray burst prompt emission light curves.
Gao H., Zhang B.-B., Zhang B.
<Astrophys. J., 748, 134 (2012)>
=2012ApJ...748..134G 2012ApJ...748..134G
ADC_Keywords: Gamma rays
Keywords: gamma-ray burst: general
Abstract:
Gamma-ray bursts (GRBs) have variable light curves. Although most
models attribute the observed variability to one physical origin
(e.g., central engine activity, clumpy circumburst medium, or
relativistic turbulence), some models invoke two physically distinct
variability components. We develop a method, namely, the stepwise
filter correlation method, to decompose the variability components in
a GRB light curve. Based on a low-pass filter technique, we
progressively filter the high-frequency signals from the light curve,
and then perform a correlation analysis between each adjunct pair of
filtered light curves. Our simulations suggest that if a mock light
curve contains a "slow" variability component superposed on a rapidly
varying time sequence, the correlation coefficient as a function of
the filter frequency would display a prominent "dip" feature around
the frequency of the slow component. Through simulations, we
demonstrate that this method can identify significant clustering
structures of a light curve in the frequency domain, and we prove that
it can catch superposed signals that are otherwise not easy to
retrieve based on other methods (e.g., the power density spectrum
analysis method). We apply this method to 266 Burst and Transient
Source Experiment bright GRBs. We find that the majority of the bursts
have clear evidence of such a superposition effect. We perform a
statistical analysis of the identified variability components and
discuss the implications for GRB physics.
Description:
We select 266 bright GRBs detected by BATSE (Kaneko et al. 2006,
Cat. J/ApJS/166/298), whose light curve data and T90 values are
publically available from the online database at
http://heasarc.gsfc.nasa.gov/docs/cgro/batse/
The total 266 bursts could be grouped into four categories based on
both their light curves and SFC (stepwise filter correlation method)
curves. The first category, the "Good sample" includes 117/266 (44.0%)
of the bursts. They clearly show at least one dip in the SFC curve.
Checking back along the light curves, one can usually find one or more
pulses with the identified characteristic frequencies. Superposed on
the identified slow component, there are always more rapid variability
features. This clearly suggests a superposition of at least two
variability components in the light curves (See section 3).
File Summary:
--------------------------------------------------------------------------------
FileName Lrecl Records Explanations
--------------------------------------------------------------------------------
ReadMe 80 . This file
table1.dat 75 126 Characteristic timescales identified in BATSE
bright gamma-ray bursts
--------------------------------------------------------------------------------
See also:
IX/20 : The Fourth BATSE Burst Revised Catalog (Paciesas+ 1999)
J/ApJ/744/141 : Shapes of GRB light curves (Bhat+, 2012)
J/ApJS/166/298 : Bright BATSE gamma-ray bursts spectral cat. (Kaneko+, 2006)
J/ApJS/126/19 : BATSE gamma-ray burst spectral catalog. I. (Preece+, 2000)
http://heasarc.gsfc.nasa.gov/docs/cgro/batse/ : BATSE NASA home page
Byte-by-byte Description of file: table1.dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 7 A7 --- Name GRB or 4B designation (YYMMDDA) (1)
9- 13 F5.1 s T90 [6/319] Burst T90 duration from
BATSE (Cat.IX/20)
15- 19 F5.1 s T1 [1/226] Characteristic timescale corresponding
to the first dip (2)
21- 24 F4.2 --- s1 [0/1] significance of first dip (3)
26- 29 F4.2 --- c1 [0.9/1] T1 c parameter (3)
31- 34 F4.1 s T2 [0/89]?=0 Characteristic timescale corresponding
to the second dip (2)
36- 40 F5.3 --- s2 [0/1]? significance of second dip (3)
42- 45 F4.2 --- c2 [0.9/1]? T2 c parameter (3)
47- 50 F4.1 s T3 [0/28]?=0 Characteristic timescale corresponding
to the third dip (2)
52- 55 F4.2 --- s3 [0/1]? significance of third dip (3)
57- 60 F4.2 --- c3 [0.9/1]? T3 c parameter (3)
62- 64 F3.1 s T4 [0/9]?=0 Characteristic timescale corresponding to
the fourth dip (2)
66- 70 F5.3 --- s4 [0/1]? significance of fourth dip (3)
72- 75 F4.2 --- c4 [0.9/1]? T4 c parameter (3)
--------------------------------------------------------------------------------
Note (1): Within sample I (good sample), ∼25.6% bursts show just one dip
in the low-frequency regime (first part of the table). The remaining
bursts (∼74.4%) in sample I show more than one dip. For each dip we
try to identify the corresponding component in the light curve. See
section 3 for further explanations.
910321 is very likely a misprint for 940321; corrected at CDS.
Note (2): Since there is no strict periodicity in the light curves, the
timescales of all the components we have identified are rough values,
and we have rounded them to the nearest 0.5s.
Note (3): In order to quantitatively delineate the significance and
confidence level of each dip, we define two parameters. The
significance parameter, s, delineates the depth/shallowness of a dip
in the stepwise filter correlation (SFC) Ri-fc,i curve. See
equation (2). Next, we define a confidence level parameter, c, based
on Monte Carlo simulations (c≥0.9 is regarded as a high confidence
level). See section 3 for further explanations.
--------------------------------------------------------------------------------
History:
From electronic version of the journal
(End) Emmanuelle Perret [CDS] 22-Oct-2013