| Abstract
| - Context. We present a method for determining the background of the gamma-ray bursts (GRBs) of the Fermi Gamma-ray Burst Monitor (GBM) using the satellite positional information and a physical model. Since the polynomial fitting method typically used for GRBs is generally only indicative of the background over relatively short timescales, this method is particularly useful in the cases of long GRBs or those that have autonomous repoint request (ARR) and a background with much variability on short timescales. Aims. Modern space instruments, like Fermi, have some specific motion to survey the sky and catch gamma-ray bursts in the most effective way. However, GBM bursts sometimes have highly varying backgrounds (with or without ARR), and modelling them with a polynomial function of time is not efficient - one needs more complex, Fermi-specific methods. This article presents a new direction dependent background fitting method and shows how it can be used for filtering the lightcurves. Methods. First, we investigate how the celestial position of the satellite may have influence on the background and define three underlying variables with physical meaning: celestial distance of the burst and the detector’s orientation, the contribution of the Sun and the contribution of the Earth. Then, we use multi-dimensional general least square fitting and Akaike model selection criterion for the background fitting of the GBM lightcurves. Eight bursts are presented as examples, of which we computed the duration using background fitted cumulative lightcurves. Results. We give a direction dependent background fitting (DDBF) method for separating the motion effects from the real data and calculate the duration ( T90, T50, and confidence intervals) of the nine example bursts, from which two resulted an ARR. We also summarize the features of our method and compare it qualitatively with the official GBM Catalogue. Conclusions. Our background filtering method uses a model based on the physical information of the satellite position. Therefore, it has many advantages compared to previous methods. It can fit long background intervals, remove all the features caused by the rocking behaviour of the satellite, and search for long emissions or not-triggered events. Furthermore, many parts of the fitting have now been automatised, and the method has been shown to work for both sky survey mode and ARR mode data. Future work will provide a burst catalogue with DDBF.
|
