Direction Reconstruction of Radio Signals in Neutrino Detectors in Ice

One of the major outstanding questions in astroparticle physics is the origin of ultra-high energy cosmic rays. Because charged cosmic rays are deflected by (inter)galactic magnetic fields, they do not point back to their sources, and photons are scattered by the low-energy background radiation that is present throughout the universe. This leaves neutrinos as the prime messenger to identify the sources of ultra-high energy cosmic rays. Due to their very low expected flux, any experiment aiming to detect ultra-high energy neutrinos needs to have a very large effective volume. This requirement can be fulfilled by detecting the radio emission from particle cascades in ice, which can travel kilometre-scale distances before being attenuated, allowing for a sparser and therefore more cost-effective detector array. The Radio Neutrino Observatory Greenland (RNO-G) is the first full-scale in-ice radio detector, and the proposed successor to the optical IceCube telescope, IceCube-Gen2, is expected to include a 500-square-kilometre in-ice radio array. These experiments will be the first in-ice radio observatories with a realistic chance of detecting ultra-high energy neutrinos. However, in order to identify these neutrinos and their sources, one has to be able to distinguish them from background signals as well as reconstruct their arrival direction from radio data. This thesis contains contributions to the reconstruction algorithms for wind-induced impulsive events, cosmic-ray air showers and neutrinos. The first two serve to better understand two of the more common sources of background signals for in-ice radio detectors, whereas the latter is crucial to identify the sources of ultra-high energy neutrinos. The uncertainties on the reconstructed neutrino direction turn out to be strongly asymmetric, necessitating a two-dimensional treatment.