Constraining the Potential of the Milky Way using Stellar Streams

Sammanfattning: The orbital approximation for stellar streams is the assumption that all stars in a stream can be described as following the same orbit. In this thesis, I evaluate this approximation as a method for constraining the potential of the Milky Way. I wrote a program in MATLAB which integrates the orbits of test particles in a potential model which resembles that of the Milky Way's halo and disc. The group of particles are set off with a velocity and position dispersion to act like a typical satellite. By varying the satellite's initial position, velocity and angle of the velocity relative to the disc, a large variety of different streams were created. The formed streams were treated as real observations, to determine how well the method works on actual streams. Using the position and velocity of the central particle in the satellite, orbits were fitted to the stream by altering the disc and halo masses and calculating the log-likelihood of each orbit. I try to constrain the mass values using a 90% confidence region in a 2D log-likelihood plot. The error in the orbital approximation is clearly visible when plotting the central orbit together with the simulated stream stars. No specific initial parameter (of the ones changed) gave a predictable confidence region, meaning that the approximation fails whatever the details of the stream models. It is found that most streams do not produce a confidence region that contain the masses of the potential they were actually created in. The percentage of confidence regions which had the actual masses inside them were 22.90%, which is very far from the 90% expected (from the confidence region) if the the orbital approximation was accurate. Only using streams with confidence regions entirely in the positive mass region gave a percentage of 12.50%. This means that past studies which used the orbit approximation are likely to have found incorrect answers regarding the shape and mass of the Milky Way.