The capability of the Fermi Gamma-ray Burst Monitor (GBM) to localize gamma-ray bursts (GRBs) is evaluated for two different automated algorithms: the GBM Teams RoboBA algorithm and the independently developed BALROG algorithm. Through a systematic study utilizing over 500 GRBs with known locations from instruments like Swift and the Fermi LAT, we directly compare the effectiveness of, and accurately estimate the systematic uncertainty for, both algorithms. We show simple adjustments to the GBM Teams RoboBA, in operation since early 2016, yields significant improvement in the systematic uncertainty, removing the long tail identified in the systematic, and improves the overall accuracy. The systematic uncertainty for the updated RoboBA localizations is $1.8^circ$ for 52% of GRBs and $4.1^circ$ for the remaining 48%. Both from public reporting by BALROG and our systematic study, we find the systematic uncertainty of $1-2^circ$ quoted in GCN circulars for bright GRBs localized by BALROG is an underestimate of the true magnitude of the systematic, which we find to be $2.7^circ$ for 74% of GRBs and $33^circ$ for the remaining 26%. We show that, once the systematic uncertainty is considered, the RoboBA 90% localization confidence regions can be more than an order of magnitude smaller in area than those produced by BALROG.
تحميل البحث