The NGRG paper is all about making precise, conservative estimates of GRB observation and detection rates based on actual Swift on-orbit performance. For this reason, we begin with an analysis of Swift detection and observation rates which we hope will be taken as authoritative.

Rate Definitions - Swift can respond to alerts from other instruments, and can observe bursts found post-facto in "ground" analysis. Swift can observe GRBs but not detect them in XRT and UVOT. There are many combinations and one has to take care what definition of measurement and detection one uses (don't forget about upper limits vs. detections!). Rates and other measurements should be objective and the value should not depend on who makes the measurement; for objective rates specify your definition carefully (read my footnotes).

Measurement Periods- At one point, Swift lost a gyro and had some additional restrictions on what it could follow up on. There have been several periods of non-operation due to everything from systems failures (usually fixed eventually) to moving the Swift offices. So be careful to specify the measurement period if you expect to get the same rate. In particular, just after Swift was turned on, the BAT detection rate changed due to fiddling with the detection algorithms and ramping up operations. During the first couple years of operation, the UVOT exposure sequence changed, and so for e.g. the standard "FC" exposure the sensitivity changed. Also, the sensitivity of the UVOT detector is decreasing somewhat with time. Below, you will see that we carefully select the periods for consistent performance.

Basic Swift Rates

Appendix Table A1  Swift GRB Measurement Rates

Start Time	End Time	Parameter	Value	Rate (yr.–1)	Note
2005 Jan 1.0	2013 Jan 1.0	BAT detections	736	92.0	a
2005 July 1.0	2013 Jan 1.0	UVOT observations	600	80.0	b
2005 July 1.0	2013 Jan 1.0	" ", no ground anal. events	581	77.5	b
2005 July 1.0	2013 Jan 1.0	UVOT detections	228	30.4	c
2005 July 1.0	2013 Jan 1.0	" ", no ground anal. events	224	29.9	c
Source: http://heasarc.nasa.gov/docs/swift/archive/grb_table/ a. The first publicly released burst was 041217; we chose some short time after that such that operations would be relatively stable.  All bursts, including those detected in ground analysis, are included. b. The first publicly released UVOT observation was of 050124; we chose some short time after that such that operations would be relatively stable.  c. The first publicly released UVOT detection was 050318; we chose some short time after that such that operations would be relatively stable.

Rates with Selections

Here is an analysis with more specific selections and, you will note, get slightly different answers:

• Swift Data Selection
  We wanted to select a data set with uniform X-ray and UVOT early observations.  Starting around 2006 May, initial ("finding chart") exposures were mostly around 100 s in W. After 2008, this exposure was changed to around 147 seconds. (See Notes to Table 3 below). 

• X-ray Data Set: "Early Sample X-ray Super-Sample"
  Motivated by the desire for shorter exposures for our optical reductions, we selected 2006 May 2 – 2008 Oct 7, and analyzed 224 BAT light curves in this period. See the Column "NBAT Selected" in Table 3 just below.
  SNR vs. Detecting Area Results To Be Posted.

Table 3 from Grossan+14: Swift GRB Sub-Samples and W Detection Rates

		<t_mid> (s)	texp (s)	NBAT(1) Selected (Total)	NUVOT(2) Early W Detections	UVOT Early W Detection Rate(3) (% [yr–1])
*060502a-081007	t_mid < 170s	144	~ 98	207 (224)	44	21.2 [18.1]
081008-121229a	t_mid < 220s	179	~147	357 (379)	67	18.8[15.8]
(*) The UVOT Early Brightness Distribution Sample (1) NBAT Selected includes only bursts detected by our simple algorithm, and only real-time rate triggers, as other types of triggers do not permit rapid optical follow-up by UVOT.  (The selection logic is given by: (not an image trigger) and (not a ground processing detection) and (not a failure of our trigger algorithm).)  (2) NUVOT gives the number of detections for the early sample selection (tmid ≤ 170 s and BAT selected; see following note).   (3) For rate calculation, the time period for the first sample is 2006 May 2.0 to 2008 Oct. 8.  As there were bursts on the days before and after, and the number of days in the sample time is many hundred, we did not strive for greater than 1 day accuracy.  For the second time period, the begin time is 2008 Oct. 8.0, given a burst the day before; the end time was the "blind" pre-selected, date 2013 Jan 1.0.

 

 

Early UVOT Brightness Distribution Sample
Selection Criteria: W filter/ ~100 s exposure / t_mid < 170 s
We selected all the UVOT W exposure detections 2006 May 2 – 2008 Oct 7with t_mid < 170 s and t_exp near 100 s (i.e. < 125 s), and got the resulting W values from the HESARC site. The summary statistics are in Table 3, above. The Distribution of brightness during on this these first observations at "early" times is given in the figure below. The histogram data are here. Please reference Grossan, Kumar, Perley, and Smoot 2014, for now http://arxiv.org/abs/1402.4856 . Please use these data as much as possible to benefit understanding of GRBs; please ask me to re-format or provide more data for you - I would love to collaborate with anyone interested in my data.



 

 

 



Dr. Bruce Grossan

bruce [at] singu[dot]lbl[dot]gov [anti-spam format address]

510.486.5489
Fax 510.486.7149

---


Mail comments to bruce [at] singu[dot]lbl[dot]gov [anti-spam format address]
Last Update: 2014 Sep 9