1:
2:
3:
4:
5: %\documentclass[manuscript]{aastex}
6:
7: \documentclass[12pt,preprint]{aastex}
8: \linespread{2.0}
9:
10:
11:
12:
13:
14:
15: \shorttitle{Temporal distribution of Cosmological Transients} \shortauthors{Howell et al.}
16:
17:
18:
19: \begin{document}
20:
21: \title{An improved method for estimating source densities using the temporal distribution of Cosmological Transients}
22:
23:
24: \author{E. Howell\altaffilmark{1}, D. Coward\altaffilmark{1}, R. Burman\altaffilmark{1} and D. Blair\altaffilmark{1}}
25: \affil{School of Physics, University of Western Australia, Crawley WA 6009, Australia}
26: \email{ejhowell@physics.uwa.edu.au} \altaffiltext{1}{School of Physics, University of Western Australia, Crawley WA
27: 6009, Australia}
28:
29: \newpage
30:
31: \begin{abstract}
32: It has been shown that the observed temporal distribution of transient events in the cosmos can be used to constrain
33: their rate density. Here we show that the peak flux--observation time relation takes the form of a power law that is
34: invariant to the luminosity distribution of the sources, and that the method can be greatly improved by invoking time
35: reversal invariance and the temporal cosmological principle. We demonstrate how the method can be used to constrain
36: distributions of transient events, by applying it to \emph{Swift} gamma-ray burst data and show that the peak
37: flux--observation time relation is in good agreement with recent estimates of source parameters. We additionally show
38: that the intrinsic time dependence allows the method to be used as a predictive tool. Within the next year of
39: \emph{Swift} observation, we find a 50\% chance of obtaining a peak flux greater than that of GRB 060017 -- the highest
40: \emph{Swift} peak flux to date -- and the same probability of detecting a burst with peak flux $>$ 100 photons
41: $\mathrm{s}^{-1} \mathrm{cm}^{-2}$ within 6 years.
42: \end{abstract}
43:
44: %% Keywords should appear after the \end{abstract} command. The uncommented
45: %% example has been keyed in ApJ style. See the instructions to authors
46: %% for the journal to which you are submitting your paper to determine
47: %% what keyword punctuation is appropriate.
48:
49: \keywords{gamma-rays: bursts --gravitational waves -- cosmology: miscellaneous}
50:
51: %% From the front matter, we move on to the body of the paper.
52: %% In the first two sections, notice the use of the natbib \citep
53: %% and \citet commands to identify citations. The citations are
54: %% tied to the reference list via symbolic KEYs. The KEY corresponds
55: %% to the KEY in the \bibitem in the reference list below. We have
56: %% chosen the first three characters of the first author's name plus
57: %% the last two numeral of the year of publication as our KEY for
58: %% each reference.
59:
60:
61: %% Authors who wish to have the most important objects in their paper
62: %% linked in the electronic edition to a data center may do so by tagging
63: %% their objects with \objectname{} or \object{}. Each macro takes the
64: %% object name as its required argument. The optional, square-bracket
65: %% argument should be used in cases where the data center identification
66: %% differs from what is to be printed in the paper. The text appearing
67: %% in curly braces is what will appear in print in the published paper.
68: %% If the object name is recognized by the data centers, it will be linked
69: %% in the electronic edition to the object data available at the data centers
70: %%
71: %% Note that for sources with brackets in their names, e.g. [WEG2004] 14h-090,
72: %% the brackets must be escaped with backslashes when used in the first
73: %% square-bracket argument, for instance, \object[\[WEG2004\] 14h-090]{90}).
74: %% Otherwise, LaTeX will issue an error.
75:
76: \newpage
77:
78: \section{INTRODUCTION}
79:
80: The brightness distribution of cosmological sources is conventionally used to constrain the luminosity function of the
81: sources, their evolution in density \citep{Peebles} and, for transient sources, their rate density
82: \citep{schmidt01,sethi01,totani97}. This method is applicable both to long-lived sources such as galaxies and to
83: transient events such as supernovae and gamma-ray bursts (GRBs). Estimates are obtained by fitting the number --
84: brightness distribution to models that include luminosity, source density and evolution effects. In the case of
85: transient events an additional parameter is available -- the event arrival times.
86:
87:
88: The temporal distribution of transient astrophysical populations of events has been described by the `probability event
89: horizon' (PEH) concept of \cite{coward05a}. This method establishes a temporal dependence by noting the occurrences of
90: successively brighter events in a time series. By utilizing the fact that the rarest events will preferentially occur
91: after the longest observational periods, it produces a data set with a unique statistical signature.
92:
93: Here we show that a well-defined observation-time dependence is an intrinsic feature of the source distribution of
94: events. Using Swift GRB data we demonstrate how this property can be used to constrain source distributions. We start
95: by presenting an analytical derivation of the peak flux--observation time relation, $P(T)$, for sources which are
96: uniformly distributed in Euclidean space and then describe its extension to cosmological models
97: (\S\ref{section_derivation}). We derive a simple power-law relation for $P(T)$ that is invariant to the luminosity
98: distribution of events.
99:
100: We then utilize the PEH technique to show how $P(T)$ data can be extracted from a distribution of peak fluxes (\S
101: \ref{section_enhancedPEH}). We show that the PEH method can be greatly improved by invoking time reversal invariance
102: and the temporal cosmological principle: for time scales that are short compared to the age of the Universe, a
103: distribution of independent events is invariant with respect to temporal direction and there is nothing special about
104: the time when we switch on our detector.
105:
106: As a test, we apply the $P(T)$ relation to \emph{Swift} long-GRB data. We demonstrate that the technique can be used as
107: a probe of the event rate density and luminosity distribution of the sources (\S\ref{section_FittingTool}), and as a
108: tool to predict the likelihood of future high peak flux events (\S\ref{section_PredictiveTool}).
109:
110:
111: \vspace{-4.0mm}
112: \section{THE PEAK FLUX -- OBSERVATION TIME RELATION}
113: \label{section_derivation}
114: %\subsection{Log P -- Log T relation for a Euclidean Universe}
115:
116: In this paper we will define an \emph{event} to be an astrophysical transient occurrence with a duration much less than
117: the period of observation. Examples are GRBs and gravitational wave burst sources such as coalescing compact binaries
118: or core-collapse supernovae.
119:
120: Consider a distribution of events defined within a Euclidean space by an event rate density $r_{0}$ and a luminosity
121: function\footnote{We use here the luminosity function for GRB sources, $\phi(L)$, which includes a normalization
122: constant to ensure that it integrates to unity over the range of source luminosities. This means that $\phi(L)$ has
123: units of inverse luminosity--see for example, \cite{pm01}.} $\phi(L)$ $(L_{\mathrm{min}}\le L \le L_{\mathrm{max}})$.
124: The observed peak flux, or `brightness', distribution of events over an observation time $T$ is a convolution of the
125: radial distribution of the sources and their luminosity function. For peak fluxes (photons
126: $\mathrm{cm}^{-2}\mathrm{s}^{-1}$) between $P$ and $P + \mathrm{d}P$:
127:
128: \begin{equation}
129: \mathrm{d}N(P) = 4\pi T \int_{L_{\mathrm{min}} }^{L_{\mathrm{max}} } \phi(L)dL
130: \hspace{1mm}r_{0} r^{2}dr , \label{eq_diff_source_count_euc}\\
131: \end{equation}
132:
133: \noindent with $r=(L/4\pi P)^{1/2}$. The total number of events observed in time $T$ with a peak flux greater than $P$
134: is given by :
135:
136:
137: %\begin{eqnarray}
138: %\begin{align}
139: \begin{equation}
140: % \hspace{5mm} N(> F) = N(P')\mathrm{d}P'\nonumber\\
141: N(> P) = T \Delta\Omega \int_{L_{\mathrm{min}} }^{L_{\mathrm{max}} } \phi(L)dL \int_{0}^{\sqrt{ L/4{\pi}
142: P}}\!\!\!\!r_{0}\hspace{0.5mm}r^{2}dr\,, \label{eq_NF1}
143: %\end{align}
144: %\end{eqnarray}
145: \end{equation}
146:
147: \noindent where the average solid angle covered on the sky has been accounted for by $\Delta\Omega /4 \pi$. The upper
148: limit in the integration over $r$ is the maximum distance for which an event with luminosity $L$ produces a peak flux
149: $P$.
150:
151: For $r_{0}$ and $\phi(L)$ independent of position, integrating over the radial distance yields:
152:
153: \begin{equation}
154: N(> P) = \frac{T r_{0}\hspace{0.5mm} \Delta\Omega/4\pi }{3 \hspace{0.5mm}\sqrt{4 \pi}}\hspace{0.5mm}
155: P^{-3/2}\hspace{0.5mm}\int_{L_{\mathrm{min}} }^{L_{\mathrm{max}} } \phi(L)L^{3/2} dL\,.
156: \label{eq_NF2}
157: \end{equation}
158:
159: \noindent This is the familiar log $N$--log $P$ relation, \mbox{$N(>P) \propto P^{-3/2}$}, a power law independent of
160: the form of the luminosity function \citep{horack94}.
161:
162: To introduce the temporal distribution of events, we note that, as the events are independent of each other, the
163: individual events will follow a Poisson distribution in time. Therefore, the temporal separation between events will
164: follow an exponential distribution, defined by a mean event rate $R(r) = r_{0}(4/3)\pi r^{3}$ for events out to $r$.
165: The probability for at least one event $>P$ to occur in a volume bounded by $r$ during an observation time $T$ at
166: constant probability $\epsilon$ is given by:
167:
168: \begin{equation}\label{eq_peh}
169: \mathcal{P}(n \ge 1;R(r),T)= 1 - e^{R(r)T} = \epsilon \,.
170: \end{equation}
171:
172: \noindent For this equation to remain satisfied with increasing observation time:
173:
174: \begin{equation}\label{eq_eps}
175: N(>P) = R(r)T = |\mathrm{ln}(1 - \epsilon)|. \\
176: \end{equation}
177:
178: Equations (\ref{eq_NF2}) and (\ref{eq_eps}) for $N(>P)$ combine to give the relation for the evolution of brightness as
179: a function of observation time:
180:
181:
182: \begin{equation}
183: P(T)\! =\!\! \left(\frac{r_{0}\hspace{0.5mm}\Delta\Omega /4\pi\hspace{0.5mm} }{3 \hspace{0.5mm}\sqrt{4
184: \pi}\hspace{0.5mm}|\mathrm{ln}(1 - \epsilon)|}\right)^{\!\!2/3}\!\! \left[\int_{L_{\mathrm{min}}
185: }^{L_{\mathrm{max}}}\!\!\!\!\!\!\!\phi(L)L^{3/2}dL \right]^{2/3}\!\!\!\!\!\!T^{2/3}\!\hspace{0.5mm}.
186: \label{eq_logPlogT_euc}\\
187: \end{equation}
188:
189:
190: This relation shows that for a simple Euclidean geometry, a log $P$--log $T$ distribution will have a slope of 2/3,
191: independent of the form of the luminosity function. One can consider that changes in $r_{0}$ create a horizontal offset
192: in the log $P$--log $T$ distribution, while changes in the integrated luminosity create a vertical offset. However, the
193: slope is fixed by the 3-D Euclidean geometry.
194:
195:
196: We can use the log $P$--log $T$ relation to produce curves defining the probability, $\epsilon$, of obtaining some
197: value of peak flux, $P$, within an observation-time, $T$, for a given $r_{0}$ \mbox{and $\phi(L)$}.
198:
199: For a cosmological distribution of sources, equation (\ref{eq_logPlogT_euc}) must be modified to allow for cosmic
200: evolution. A standard Friedman cosmology can be used to define a differential event rate, $\mathrm{d}R(z)$, in the
201: redshift shell $z$ to $z + \mathrm{d}z$. The luminosity and flux will be related through $z$ by a luminosity distance
202: $d_{\mathrm{L}}^{2}(z)$ (see for example \cite{coward05a} or \cite{pm01}). In this case, solving equation
203: (\ref{eq_eps}) numerically, with $P = L/4\pi d_{\mathrm{L}}^{2}(z)$, will yield the cosmological log $P$--log $T$
204: relation.
205:
206:
207:
208:
209:
210: \vspace{-4.0mm}
211: \section{AN ENHANCED PEH FILTER}
212: \label{section_enhancedPEH}
213:
214: To utilize the time domain, we use the probability event horizon (PEH) filter of \cite{coward05a} to produce $P(T)$
215: time data. The PEH filter is a tool that exploits the temporal information encoded in a time series of transient events
216: and works by recording successively brighter events in a time series. \cite{howell07} demonstrated that the unique
217: statistical signature of events filtered in this way could be exploited to obtain rate estimates of transient events.
218: However, the significant probability of a bright event occurring early in an observational period meant that only a
219: small fraction of data was used by the method. As a result, large uncertainties were obtained in the estimates. There
220: are however, two ways in which the amount of usable data can be increased.
221:
222: Firstly, the temporal cosmological principle implies that the PEH signature of a transient population of events is
223: independent of when a detector is switched on. Secondly, time reversal invariance allows the PEH filter to be applied
224: to a data set in both temporal directions. Thus, a time series of events can be treated as a closed loop which can be
225: interrogated in both directions. The observational period is now defined as the total length of the loop. The start
226: time for the PEH analysis is now arbitrary so, without loss of generality, we can choose any start time. This allows
227: the PEH filter to be applied in such a way that the brightest event can be set as the final event in a series. This
228: ensures that the PEH filter is applied to the full data stream and the process can be repeated in each direction,
229: increasing the quantity of PEH data. We refer to these techniques as `from max' plus `time reversal' (FMTR). We show
230: below how FMTR increases the PEH sample and significantly improves the statistical resolution when applied to the
231: \emph{Swift} data.
232:
233:
234:
235:
236:
237: \section{APPLICATION TO \emph{SWIFT} DATA} \label{section_FittingTool}
238: In this section, we will apply the log $P$--log $T$ relation to a cosmological population of long GRBs. To account for
239: the event rate and luminosity function, we will use estimates based on recent studies. Using the FMTR method, we will
240: extract a time-dependent sample of GRB peak fluxes from the \emph{Swift} data, and
241: demonstrate how it can be constrained by a log $P$--log $T$ fit.\\
242: \indent For our long-GRB peak flux sample we use data recorded by the \emph{Swift}\footnote{This data can be obtained
243: from the \emph{Swift} website http://swift.gsfc.nasa.gov/docs/swift/archive/grb\_table.html} satellite between 2004
244: December and 2007 April. We consider only bursts with confirmed peak fluxes detected within the 15--150 keV band of the
245: Burst Alert Telescope (BAT) and with $T_{90} \ge 2\hspace{0.5mm}s$ (a $T_{90}$ duration is the interval in which a
246: signal contains 90\% of its total observed counts). The total sample consists of
247: 190 peak fluxes.\\
248: \indent Figure \ref{plotone}A displays the \emph{Swift} peak flux distribution of long GRBs as a time series. It is
249: apparent that as observation time increases, there is an greater probability of a bright event. By extracting
250: successively brighter events as a function of observation time, the PEH filter samples events from the low probability
251: tail of the distribution. The strong brightness--time dependence
252: of these events creates a unique statistical signature which can be modeled by the log $P$ -- log $T$ relation.\\
253: \indent In Table 1 we show the PEH filtered data. It is apparent how the time intervals between successive events
254: increase with observation time. This is a result of a progressive sampling of the rarer events of the
255: distribution.\\
256: \indent To apply the log $P$--log $T$ relation to the filtered data we must first set up a model to account for the
257: source rate evolution and luminosity distribution. We use a model from the recent study of \cite{guetta07}. They employ
258: a `flat-$\Lambda$' cosmological model and \mbox{$H_{0}=65$ km s$^{-1}$ Mpc$^{-1}$} for the Hubble parameter at the
259: present epoch. For the isotropic luminosity function of GRB peak luminosities, $\phi(L)$, they use a broken power law
260: form based on the work of \cite{schmidt01}. Assuming that the rate of GRBs traces the global star formation history of
261: the Universe, they employ a number of different star formation rate models. For each, they determine best fitted values
262: for the luminosity function and event rate density. For this study, we use their model (i) parameters, which are based
263: on the SF2 star formation rate model of \cite{pm01}. These parameters include fitted values for the luminosity function
264: and a local event rate density $r_{0} = 0.1\hspace{0.5mm} \mathrm{Gpc}^{-3} \mathrm{yr}^{-1}$. To account for the
265: average solid angle covered on the sky by \emph{Swift}, we use a value of $\Delta\Omega =
266: 1.33$ \citep{Band03}.\\
267: \indent Figure \ref{plotone}B shows a log $P$-- log $T$ fit to the PEH filtered data (shown by squares), using the
268: fitting parameters of \cite{guetta07}: i.e. there are no free parameters in the comparison between theory and
269: observation. We define a 90\% confidence band -- shown by the shaded area -- corresponding to the $\epsilon = 95\%$
270: (top) and $\epsilon = 5\%$ (bottom) probabilities of detecting at least one event within an observation time $T$. We
271: see that the data is well constrained. The fit shows that by using only a small sample of the brightest events, it is
272: possible to extract the geometrical signature of the source population and to test estimates of the luminosity
273: function and rate density of events.\\
274: \indent The dashed lines of Figure \ref{plotone}B show the 90\% confidence band corresponding to the Euclidean model
275: using equation (\ref{eq_logPlogT_euc}). We see that two of the first few events lie outside the Euclidean curves but
276: are constrained by the cosmological model. These events, occurring at early observation times, most likely result from
277: sources at large cosmological distances. The very bright events at late observation times are more probable -- it is
278: apparent that the Euclidean and cosmological curves begin to converge in this regime. To account for non-uniformly
279: distributed sources, the method could be refined to take into account the spatial distribution of potential host
280: galaxies.
281:
282: \indent To test the power-law dependence in equation (\ref{eq_logPlogT_euc}), we have performed least-squares fitting
283: to the PEH filtered data using the Euclidean log $P$-- log $T$ curve as a linear regression model. By setting the power
284: as a single free parameter, we obtained a value of $0.67 \pm 0.02$, confirming the 2/3 slope derived in section
285: \ref{section_derivation}.
286:
287: \vspace{-4.0mm}
288: \section{A PREDICTIVE APPLICATION OF THE LOG P -- LOG T RELATION} \label{section_PredictiveTool}
289:
290: %\begin{enumerate}\item
291:
292: Figure \ref{plottwo} illustrates how the log $P$--log $T$ relation can be used as a predictive tool. By mapping the
293: temporal evolution of detection probability, the maximum brightness of future events can be constrained
294: \citep{coward05b}. As in Fig. \ref{plotone}B, the shaded areas show the log $P$--log $T$ detection confidence bands
295: corresponding to different probability values (shown in the legend). The current \emph{Swift} observation time (398
296: days) is shown by the vertical solid line. For predictive applications it is essential that the true temporal sequence
297: of events is retained. Therefore, rather than using the FMTR technique, we apply the unmodified PEH filter from the
298: time of the first event. Comparing with Fig. \ref{plotone}B, we see that the FMTR method has increased the sample by
299: 220\%, of which 80\% is gained by incorporating time reversal. Using 100 Monte-Carlo simulations, we find a mean
300: fractional increase in data of $\sim 35\%$, compared to the unmodified PEH filter. The larger than expected filter
301: output using \emph{Swift} data, implies that the FMTR method can be further optimally tuned.
302:
303: The plot shows that after 3 days of operation, there was a $<5\%$ probability of detecting an event with peak flux
304: equal to the first in the PEH sample, GRB 041123. The Poisson probability of detecting at least one event within a Gpc
305: at this time is $\sim 2 \times 10^{-6}$. This implies that this event occurred at a considerable cosmological distance.
306: The event GRB 050525A is the brightest long GRB with a secure redshift, $z=0.606$ -- the Poisson probability of at
307: least one event within this volume is 35\%. The next brightest event, GRB 060017, is the most intense burst, in terms
308: of peak flux, detected by Swift.
309:
310:
311: As a demonstration of the predictive nature of the the log $P$ -- log $T$ relation, Figure \ref{plottwo} shows that
312: there is a 50\% probability of obtaining an event with a peak flux greater than that of GRB 060017 within the next year
313: and an 80\% probability within 5 years.
314:
315: The curve predicts that there is a 50\% (80\%) probability of obtaining a burst with peak flux $>$ 60 photons
316: $\mathrm{s}^{-1} \mathrm{cm}^{-2}$ within 2 (8) years. To determine the feasibility of this prediction, we consider
317: again GRB 050525A, which had a peak flux of 42.3 photons $\mathrm{s}^{-1} \mathrm{cm}^{-2}$. Using this burst's
318: redshift and converting to a peak luminosity, we find that an equivalent burst would have to occur within $z \approx
319: 0.5$ to produce a peak flux of this level. The Poisson probability of at least one event at this distance within the
320: next two years \mbox{is $\sim 90\% \hspace{0.5mm}(99\%)$}. If we consider a burst with a peak flux of 100 photons
321: $\mathrm{s}^{-1} \mathrm{cm}^{-2}$, we find that an event of this peak flux is 50\% probable within 6 years. Such an
322: event would correspond to a GRB 050525A-equivalent burst occurring within $z \approx 0.4$, for which the probability is
323: 99\%.
324:
325:
326: The log $P$ -- log $T$ technique naturally uses the brightest events of a data set. As the probability of obtaining a
327: GRB afterglow increases with peak flux, the method can be used to predict the expected occurrence of events at low $z$.
328:
329:
330:
331:
332: \section{SUMMARY}
333: \label{section_conclusions}
334:
335:
336:
337:
338:
339: We have provided a clear demonstration of the log $P$--log $T$ relation by applying the PEH filter to the \emph{Swift}
340: GRB peak flux distribution. A log $P$--log $T$ model with no free parameters was fitted to filtered data confirming the
341: power law $P \propto T^{2/3}$ in the Euclidean limit; the power law is independent of the form of the luminosity
342: distribution.
343:
344:
345: The FMTR method significantly improves the PEH method, which was previously disadvantaged by using only a small
346: fraction of a data stream. We have shown that FMTR enables the PEH filter to use over 8\% of the available data, making
347: it a practical tool for cosmology.
348:
349:
350: We have shown that the PEH technique can be used as a predictive tool. Comparing observation with prediction provides
351: an additional means to test rate estimates and evaluate source parameters such as the limits of the luminosity
352: distribution.
353:
354:
355: In a future study, we intend to apply the FMTR method to both the \emph{Swift} and \emph{BATSE} GRB data. We will
356: investigate the efficiency of the method in determining constraints on the rate density and limits of the luminosity
357: function.
358:
359:
360:
361:
362:
363:
364:
365:
366: \section*{ACKNOWLEDGMENTS}
367: E Howell and D Coward are supported by the Australian Research Council.
368: % This research is part of the research program of the Australian Consortium for Interferometric Gravitational Astronomy
369: %(ACIGA).
370: \vspace{-4.0mm}
371: \bibliographystyle{mn2e} % Choose Phys. Rev. style for bibliography
372: %\bibliography{swift_peh} % qhe.bib is the name of our database
373: \addcontentsline{toc}{chapter}{Bibliography}
374:
375: \newpage
376:
377: \begin{thebibliography}{99}
378:
379: \bibitem[\protect\citeauthoryear{Band}{2005}]{Band03}Band, D, 2003,
380: ApJ, 588, 945
381:
382: \bibitem[\protect\citeauthoryear{Coward \& Burman}{2005}]{coward05a}D.~M. Coward \& R.~R. Burman, 2005,
383: MNRAS, 361, 362
384:
385: \bibitem[\protect\citeauthoryear{Coward at al.}{2005}]{coward05b}D.~M. Coward, M. Lilley, E.~J. Howell, R.~R. Burman \& D.~G. Blair, 2005,
386: MNRAS, 364, 807
387:
388: \bibitem[\protect\citeauthoryear{Guetta \& Piran}{2007}]{guetta07}D. Guetta \& T. Piran, 2007, JCAP, in press
389: (astro-ph/0701194)
390:
391: \bibitem[\protect\citeauthoryear{Howell at al.}{2007}]{howell07}E. Howell and D. Coward and R. Burman \& D. Blair 2007,
392: MNRAS, 377, 719
393:
394: \bibitem[\protect\citeauthoryear{Horack, Emslie \& Meegan}{1994}]{horack94}J. Horack and A. Emslie \& C. Meegan, 1994,
395: ApJ, 426, L5
396:
397: \bibitem[\protect\citeauthoryear{Peebles}{1993}]{Peebles} Peebles P. J. E., 1993, Principles of Physical Cosmology,
398: Princeton Univ. Press, Princeton NJ
399:
400: \bibitem[\protect\citeauthoryear{Porciani \& Madau}{2001}]{pm01} Porciani C., Madau P., 2001,
401: ApJ, 548, 522
402:
403: \bibitem[\protect\citeauthoryear{Schmidt}{2001}]{schmidt01} Schmidt M., 2001,
404: ApJ, 552, 36
405:
406: \bibitem[\protect\citeauthoryear{Sethi \& Bhargavi}{2001}]{sethi01} Shiv Sethi \& S.~G. Bhargavi, 2001,
407: A\&A, 376, 10
408:
409: \bibitem[\protect\citeauthoryear{Totani}{1997}]{totani97} Totani T., 1997, ApJ, 486, L71
410:
411: \end{thebibliography}{}
412:
413: \newpage
414:
415:
416: \begin{table}
417: \begin{centering}
418: \begin{tabular}[scale=1.0]{cccc}
419: \hline
420: GRB &Peak Flux & Redshift & Observation Time\\
421: %GRB &Peak Flux %(ph$\hspace{0.5mm}\mathrm{s}^{\!-1} \mathrm{cm}^{\!-2})$ % & Redshift & Observation time (days)\\
422: &(photons$\hspace{0.5mm}\mathrm{s}^{-1} \mathrm{cm}^{-2})$ & & (days) \\
423: \hline
424: %060123 &0.04 & $1.099 $ & 1 \\
425: 060202 &0.5 & $ $ & 8 \\
426: 060203 &0.6 & $$ & 9 \\
427: \hspace{2.0mm}060204B &1.3 & $ $ & 10 \\
428: 060206 &2.8 & $4.045 $ & 12 \\
429: \hspace{2.0mm}060223B &2.9 & $ $ & 29 \\
430: 060306 &6.1 & $ $ & 42 \\
431: 060418 &6.7 & $1.49 $ & 84 \\
432: \hspace{2.0mm}060510A &17.0 & $ $ & 106 \\
433: 061121 &21.1 & $1.314 $ & 297 \\
434: \hspace{2.0mm}050219B &25.4 & $ $ & 506 \\
435: \hspace{2.0mm}050525A &42.3 & $0.606 $ & 602 \\
436: 060117 &48.9 & $ $ & 839 \\
437: \hline
438: %060111B & 1.4 & &4\\
439: \hspace{2.0mm}060111A & 1.72 & &4\\
440: 060110 & 1.9 & &5\\
441: 060105 & 7.5 & &10\\
442: 060603 &27.6 &2.821 &227\\
443: \hline
444: \end{tabular}
445: \caption[waveform parameters]{Data extracted from the \emph{Swift} peak flux distribution of long GRBs using the PEH
446: filter. The observation time is determined by treating the time series as a closed loop and setting the last event in
447: the series to be the brightest burst. The lower set of data was obtained by invoking time reversal. A log $P$ -- log
448: $T$ fit to this data is shown in Fig. \ref{plotone}}
449: \end{centering}
450: \label{table1}
451: \end{table}
452:
453: \newpage
454: \begin{figure}
455: \includegraphics[width=94mm]{f1.eps}
456: \caption{Panel (A) shows the \emph{Swift} peak flux distribution as a time series. It is evident that as observation
457: time increases the probability of a bright event increases. Panel (B) uses a PEH filter to extract the geometrical
458: signature of the GRB distribution (shown as squares). Assuming an event rate of $0.1 \hspace{0.5mm}\mathrm{Gpc}^{-3}
459: \mathrm{yr}^{-1}$ \citep{guetta07}, the shaded area shows the cosmology dependent log $P$--log $T$ model corresponding
460: to a (5 -- 95)\% confidence band. The equivalent model for a Euclidean geometry is shown by the dashed curves -- the
461: two outliers, which have no associated redshifts, probably result from distant cosmological events.} \label{plotone}
462: \end{figure}
463:
464: \newpage
465: \begin{figure}
466: \includegraphics[width=84mm]{f2.eps}
467: \caption{The log $P$--log $T$ relation used as a predictive tool. The successive maximum peak fluxes detected by Swift
468: since the start of operation are shown by squares. The shaded areas show detection confidence bands corresponding to
469: different probability values (shown in the legend). The current \emph{Swift} observation time (398 days) is shown by
470: the solid line. The plot shows that after 3 days of operation, there was a $<5\%$ probability of detecting an event
471: with peak flux equal to the first in the PEH sample, GRB 041123. We see that within the next year, there is a 50\%
472: chance of obtaining a peak flux greater than for GRB 060017 -- the most intense burst (in peak flux) detected by Swift
473: -- and the same probability of obtaining a burst with peak flux $>$ 100 photons $\mathrm{s}^{-1} \mathrm{cm}^{-2}$
474: within 6 years. } \label{plottwo}
475: \end{figure}
476:
477:
478: \end{document}
479: