The increasing scales of modern wind energy systems, with rotor diameters exceeding 250 m and hub heights above 150 m, introduces new challenges in understanding interactions between atmospheric dynamics and wind farm performance. This study investigates the impact of atmospheric boundary layer height (BLH) as a key parameter influencing wind farm efficiency and turbine loads. Using mesoscale simulations from the Weather Research and Forecasting (WRF) model combined with lidar measurements, we quantify BLH variability and its associated uncertainty across three representative sites in the North and Baltic Seas. A series of Computational Fluid Dynamics (CFD) simulations for a wind farm, containing 100 15 MW turbines, under varying BLH and wind speed conditions reveal significant efficiency differences linked to atmospheric stratification, with lower BLH generally reducing farm efficiency. Seasonal and site-specific climatologies highlight that Baltic Sea conditions, characterized by larger extent of low BLH conditions, lead to reduced performance compared to North Sea sites. Furthermore, we assess the influence of large-scale coherent turbulence structures on turbine loads through aeroelastic simulations of both bottom-fixed and floating configurations. The results show that low-frequency fluctuations, often absent in standard design models, increase fatigue loads within wind farms, particularly for turbines in wake-affected regions. These findings underscore the need to incorporate BLH variability and large-scale turbulence effects into engineering models for reliable performance and load predictions of next-generation offshore wind farms.