How Do Planetary Boundary Layer Schemes Perform in Hurricane Conditions: A Comparison With Large‐Eddy Simulations

Abstract

AbstractParameterizations of turbulent processes in planetary boundary layer (PBL) schemes impact tropical cyclone (TC) forecasts. Existing PBL schemes are mostly designed for low‐wind conditions, and assessing their uncertainties in hurricane conditions remains challenging, mostly due to very scarce observations. Using a recently developed framework based on large‐eddy simulations (LES), this study evaluates K‐profile parameterization (KPP) and high‐order PBL schemes in hurricane conditions. Among KPP PBL schemes, the Global Forecast System (GFS) scheme tends to produce excessively deep inflow layers with large values of eddy viscosity (Km). Opposite results are found for the Yonsei University (YSU) scheme. Using LES results as a benchmark, the performance of YSU and GFS schemes is improved by modifying the “shape parameter” such that Km is maximized closer to the surface, and by using a new definition of boundary layer height tailored to high‐wind conditions. The LES results also suggest an asymptotic mixing length of ∼40 m can improve the Louis‐type parameterizations of the YSU scheme that operates above the boundary layer. Among high‐order PBL schemes, the Mellor–Yamada–Nakanishi–Niino (MYNN) scheme produces reasonably accurate vertical profiles of eddy viscosity, turbulent stress, and boundary layer winds under different high‐wind conditions. Further analysis of MYNN supports a “three‐layer” strategy for the mixing length parameterization for TCs that represents different types of turbulent regimes. In contrast, the high‐order eddy‐diffusivity mass‐flux scheme produces excessive boundary‐layer vertical mixing and a deeper inflow layer, partly attributable to a notable overestimation of the maximum allowable mixing length in the PBL code.