Consistent directional likelihoods for IceCube realtime alerts

The IceCube Neutrino Observatory instruments a cubic-kilometer of ice at the South Pole. A primary science goal is the detection of high-energy neutrinos from astrophysical sources. In order to facilitate such discoveries, IceCube issues alerts in real time for specific categories of high-energy events, thus allowing rapid follow up by the multimessenger community. As an example, follow up observations initiated by the alert IC170922A found a flaring blazar, TXS 0506+056, in the direction of the alert, ultimately leading to the first astrophysical source detected by IceCube. However, the construction of likelihood contours for this and other historical alerts is outdated and inconsistent across per-event topologies. In this talk, several updates to the reconstruction model, based on the tremendous progress made recently towards a refined understanding of the ice, will be presented that lead to more consistent likelihood spaces across different categories of realtime alerts.