2000ApJ...535..152K

Gamma-ray bursts (GRBs) are believed to be some catastrophic event in which material is ejected at a relativistic velocity, and internal collisions within this ejecta produce the observed γ-ray flash. The angular size of a causally connected region within a relativistic flow is of the order the angular width of the relativistic beaming, γ^–1. Thus, different observers along different lines of sight could see drastically different fluxes from the same burst. Specifically, we propose that the most energetic bursts correspond to exceptionally bright spots along the line of sight on colliding shells and do not represent much larger energy release in the explosion. The energy budget for an average GRB in this model is, however, same as in the uniform shell model. We calculate the distribution function of the observed fluence for random angular-fluctuation of ejecta. We find that the width of the distribution function for the observed fluence is about 2 orders of magnitude if the number of shells ejected along different lines of sight is 10 or less. The distribution function becomes narrower if number of shells along typical lines of sight increases. The analysis of the γ-ray fluence and afterglow emissions for GRBs with known redshifts provides support for our model, i.e., the large width of GRB luminosity function is not due to a large spread in the energy release but instead is due to large angular fluctuations in ejected material. We outline several observational tests of this model. In particular, for δ-function energy distribution in explosions we predict little correlation between the γ-ray fluence and the afterglow emission as in fact is observed. We predict that the early (minutes-to-hours) afterglow would depict large temporal fluctuations whose amplitude decreases with time. Finally, we predict that there should be many weak bursts with about average afterglow luminosity in this scenario.