SpringerLink
Forum Springer Astron. Astrophys.
Forum Whats New Search Orders


Astron. Astrophys. 337, 43-50 (1998)

Previous Section Next Section Title Page Table of Contents

4. Results

The results for the all-sky search are shown in Fig. 2. It presents the cumulative number of trials (search windows) against -log[FORMULA] for the all-sky search with the four time scales we have used. The results for the MC sample are also shown. Obviously, the experimental data set is consistent with MC expectations. The set of deviations has a chance probability greater than [FORMULA]0.1. This result allows us to place an upper limit for the flux of any hypothetical GRB which may have occurred in the FOV during our observations. Depending on the declination band the resulting upper limit corresponds to two energy thresholds for every time scale.

[FIGURE] Fig. 2. Cumulative number of trials (windows) versus probability in the all-sky search. The time scale is indicated in brackets. The solid lines represent real data while the dashed lines represent the results (normalized to the real data) over a MC sample 10 (20 in the 1 hour scale) times larger than the real one. Deviations are not significant ([FORMULA]).

Assuming the same spectral index for the source and the CR flux and a steady emission during the time interval, we can estimate the integral flux upper limit with the formula (Karle et al. 1995b; Alexandreas et al. 1993b)

[EQUATION]

where [FORMULA] is the 90% CL upper limit for the integral flux, [FORMULA] and [FORMULA] are the energy thresholds for [FORMULA] and hadrons respectively, [FORMULA] is the 90% CL upper limit for the number of excess events in the source window as calculated by Aguilar-Bení tez et al. (1992) and [FORMULA] is the fraction of source events expected to fall in the source window. [FORMULA] is the known CR integral flux and is taken as [FORMULA](TeV)[FORMULA] cm-2 s-1 sr-1 (Alexandreas et al. 1993b; Burnett et al. 1990). We applied this calculation to all ON-OFF windows and determined the highest values. The resulting flux upper limits are listed in Table 1. Their uncertainties are estimated to be [FORMULA]40% through comparison with a different method of flux calculation.


[TABLE]

Table 1. 90% confidence level upper limits for the integral flux (in units of 10-8 cm-2 s-1) of any hypothetical GRB occurred in our data sample. They are tabulated depending on the time window and the declination band. Uncertainties are [FORMULA]40%.


Additional constraints on the data sample as e.g. imposed by the burst detections of BATSE, WATCH and other GRB detectors in space, can serve as a tool to reduce the full data set and thus the expected statistical fluctuations. Hence, the sensitivity is enhanced allowing detection of weaker GRBs. We therefore searched in the BATSE 3B catalog (Meegan et al. 1995) and in the WATCH catalog (Castro-Tirado 1994; Sazonov et al. 1997) for triggers which were within the FOV of AIROBICC at the time they occurred or up to 2 hours later (because of possible delayed emission). They are shown in Table 2. The sample is then reduced to events within [FORMULA] and [FORMULA]3 minutes around GRB locations. We also looked for a delayed component for two hours after the initial triggers. The size of the search region in celestial coordinates has been chosen, a priori, large enough to have the uncertainties of GRB locations into account. The length in time of the search interval (for coincident and delayed emission) has been taken according to the observations in the GeV energy range (Hurley et al. 1994).


[TABLE]

Table 2. List of GRBs observed in coincidence with satellites. [FORMULA] is the zenith angle of the GRB at the AIROBICC site.


The most significant excess in the full data set was found almost in coincidence with WATCH GRB 920925c, but from a direction 9° away from the most probable WATCH position. This observation is discussed in the next section. The results for the other three GRBs are shown in Fig. 3. It presents the cumulative number of trials against -log[FORMULA] for the four time scales used in the two search strategies (coincident and delayed emission). The results for the MC sample are shown as a dashed line. No significant deviation is seen. This, again, allows us to place an upper limit for the integral flux of these GRBs. However, we only do so for GRB 920525b because it is the only one completely covered by the data set among the three GRBs. The upper limits for the integral flux are calculated at 90% CL in the same way as shown before. The results for the coincident and delayed emission with the four time scales used here are shown in Table 3. Their uncertainties are estimated to be [FORMULA]40%.


[TABLE]

Table 3. 90% confidence level upper limits for the integral flux (in units of 10-10 cm-2 s-1) of GRB 920525b in different time scales and for two scenarios: coincident and delayed emission. Uncertainties are [FORMULA]40%.


[FIGURE] Fig. 3. Cumulative trials vs. probability in the search for the GRBs 920525b, 921118 and 930123. We show the results of the search for coincident and delayed emission. For GRB 930123 only the result for delayed emission is shown because AIROBICC started more than one hour after the burst trigger. The dashed line shows the MC results. No significant deviation appears.

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 1998

Online publication: August 6, 1998
helpdesk.link@springer.de