Mass models and the spiral arms of the Milky Way

Sammanfattning: I investigate to what extent spiral arms might affect axisymmetric Galactic models. As we need fully dynamical and high resolution Galactic evolutionary models to probe the structure of the dark matter halo, it is important that we use accurate mass models describing the MW structure. Before reaching to that level of effort, a great deal can be learned from fitting mass and kinematic models to data. Because many researchers still use models which do not include any spiral structure, they might have systematic errors from their models. I therefore compare an axisymmetric model of the Galaxy to a model with added spiral perturbations. The latter was found by adding radial and azimuthal velocity perturbations from an analytic model of spiral structure to a model of the circular velocity of the MW derived from assuming a gravitational potential with an axisymmetric bulge and disk and a logarithmic halo. Recently and accurately measured distances, proper motions and radial velocities of ∼ 100 Galactic maser sources and radial velocity data from the ISM were used to constrain the models. Using the cold gas as tracer objects in the Galaxy allows me to approximate their motions as near-circular and within the plane (z = 0). I use a Bayesian statistical analysis (with a set of priors from a variety of sources) and a Markov Chain Monte Carlo approach to investigate the parameter space of the models to arrive at a set of best fitting parameters for the two models. The resulting probability distributions for all the model parameters and the best-fitting models can then be compared and the question of spiral influence can be answered. The simple models employed in this thesis work show that including spiral perturbations does modestly change some resulting probability distributions of the model parameters and the best-fitting models. For example, I find a local DM density ρDM = 0.0096±0.001 M⊙ pc−3 for all models except for the four armed model with spiral perturbations for which ρDM = 0.00917 ± 0.001 M⊙ pc−3. However, these systematic differences are smaller than their cor- responding statistical difference given my amount of data (∼100 masers). As the amount of data increases, the statistical uncertainty will shrink (while there is no reason to expect that the systematic one will). Therefore studies of this kind will be increasingly important as the amount of data improves. I also show that choosing between a two and four armed spiral arm model gives different radial and azimuthal perturbations. For the two armed model, the radial velocity perturbation is vanishing whilst the azimuthal velocity perturbation is significant and the opposite is true for the four armed model. The point is not to derive any accurate numbers but to highlight that there is a difference for even the simplest models and simplest assumptions.