This thesis investigates turbulent wakes in stratified fluids, with a focus on the effects of the wake Reynolds number, Re. Analyses are performed on three-dimensional numerical datasets of stratified wakes produced by large-eddy simulations. Several numerical techniques are used to identify regions of distinct flow dynamics, including large-scale ‘pancake vortices,’ small-scale shear instabilities, and turbulence-emitted internal waves. By applying these techniques, it is observed that increasing wake Re leads to non-trivial effects on the wake dynamics such as modifications to the structure of the pancake vortices, prolongation of internal wave emission period, and greater longevity of small-scale shear instabilities within the anisotropic flow structures. To represent these dynamical processes in a compact way, proper orthogonal decomposition (POD) implemented by singular value decomposition is applied to the numerical data to construct reduced-order representations of the wakes. It is found that utilizing the translational symmetry in the POD implementation leads to an improvement in the convergence of POD modal energy coefficients. The effects of Reynolds number and the temporal stationarity of the data-set on the POD results are also investigated.