In this study, the offshore wind farm effects of Thanet (300 MW) in the U.K, and of Middelgrunden (40 MW) in Denmark were analyzed in a meso-scale numerical prediction model, whilst working at MeteoGroup. For the final analysis of Middelgrunden, the model is validated with measurement data from Middelgrundens Vindmøllelaug. The model uses parameterizations for the atmospheric flow, surface physics effects, and a turbine scheme based on a Drag disc concept. The scheme uses turbine data in relation to changes in the turbulent kinetic energy of the flow as it passes through the rotors. The free stream is affected by the farm through local reductions of turbulence intensity aligned in front and in-between turbine rows, and the regions affected are as high as the turbine themselves. Wind speed near the surface is found to increase inside the wind farm (a phenomena often times unexpected), transporting turbulence downstream by advection. The simulations indicate that the added turbulence form turbine rotors promotes strong vertical transport of turbulent kinetic energy up in the atmosphere, and directly above the wind farms. Further, horizontal momentum exchanges is proven to be negligible behind turbine rotors in the downstream of the farm due to the lack of proper turbulent advection. Local meteorology is affected as the air suddenly expands above the farm, increasing the height of the planetary boundary layer as a consequence. The expansion creates a vertical temperature gradient, separating cooler air masses above the rotor area, from similarly hotter air masses near the surface. The gradients obtained varied from -0.2OC to +1OC. Surface to air heat fluxes are lessen inside turbine arrays, whilst increased downstream in the farm wake. Additionally, the Local flow patterns are found to be crucial in determining the shape and a sense of direction of the wake downstream. Finally, the meso-scale model with turbine scheme is found to mimic the interaction between the turbines and the atmosphere, to some extent due to uncertainties in the horizontal gradient of turbulence intensity. Further improvement is needed to produce a precise power forecast, which also depends on the model input data, such that a focus of attention is the correct prediction of wind direction. Further, the accurate representation of the turbine layout and number of turbines per grid location are essential to detail the turbulence inside the farm, but not for analyzing meso-scale effects. Winds speed and power ratio profiles were found in agreement with measurements.