Wind farm parameterizations are essential components of mesoscale models used to assess the impact of wind farm operation on the atmospheric flow and surface climate. These models represent wind turbines as momentum sinks and sources of turbulence. We describe the theoretical basis and implementation of an improved model to enhance the Turbulent Kinetic Energy (TKE) treatment in wind farm parameterizations. The novel Latent Kinetic Energy (LKE) model integrates the tracer capabilities of the Weather Research and Forecasting (WRF) model to accurately account for wind turbine-generated TKE throughout the wakes in and downwind of wind farms. The formulation is compatible with multiple 1.5-order planetary boundary layer (PBL) schemes in the WRF model and is implemented within the explicit wake parameterization. We evaluate the LKE model against WRF large-eddy simulations with actuator-disc representations of wind turbines for an idealized 6×6 wind farm and for three 1.5-order PBL schemes. The LKE formulation improves the representation of wake turbulence, reducing normalized TKE differences relative to the large-eddy simulations to within 15 % over the wind farm region. In contrast, the Fitch wind farm parameterization with default parameter values shows 40 %. The results further show that PBL-specific calibration is required; with appropriate calibration, the LKE approach maintains normalized TKE differences within 15 % and reproduces the hub-height wind-speed deficit within 1 % of the reference.