Chang, RuiRuiChangSengers, BalthazarBalthazarSengersDörenkämper, MartinMartinDörenkämperGao, ZiqiZiqiGao2026-02-052026-02-052026https://publica.fraunhofer.de/handle/publica/50575710.1016/j.renene.2025.1250402-s2.0-105024905643Improving wind farm parameterizations in mesoscale models is essential for understanding wake interactions, and energy production. This study examined three modifications to the WRF model's Fitch wind farm parameterization (Fitch-O) - air density modification (ADM), rotor-equivalent wind speed (REWS), and axial induction modification (AIM) - and combined them into Fitch-COM to enhance real-world wind farm simulations. Validation used Supervisory Control and Data Acquisition data from a high-altitude Guazhou wind farm. Results show each modification has targeted value (Fitch-ADM for non-standard air density, Fitch-REWS for large-blade turbines, and Fitch-AIM for local-to-free-stream wind conversion) but introduces biases when used alone. The combined Fitch-COM, however, matches observations closest, showing modest but consistent gains over Fitch-O (1.7 % lower RBias, 3.0 % lower RRMSE, 2.8 % higher PCC) and a more physically reasonable framework, unlike Fitch-O (apparent accuracy from error cancelling). Extended conceptual application to inter-farm wake demonstrates both Fitch-COM and Fitch-O capture ∼8.9 %-13.2 % power deficits, with Fitch-COM simulating ∼1.0 % stronger wakes. Though limited observational data introduces uncertainty - for which we strongly call for collecting long-term data to enhance result reliability in future research - this study highlights the value of combining physics-based modifications, and supporting Fitch-COM's potential for wind farms with distinct features (e.g., high-altitude, large blades).enfalseCombined modificationFitch wind farm parameterizationReal-world atmospheric conditionsWind farm simulationWRF modeljournal article