Role of advection of parameterized turbulence kinetic energy in idealized tropical cyclone simulations

Horizontal homogeneity is typically assumed in the design of planetary boundary layer (PBL) parameterizations in weather prediction models. Consistent with this assumption, PBL schemes with predictive equations for subgrid turbulence kinetic energy (TKE) typically neglect advection of TKE. However, tropical cyclone (TC) boundary layers are inhomogeneous, particularly in the eyewall. To gain further insight, this study examines the effect of advection of TKE using the Mellor-Yamada-Nakanishi-Niino (MYNN) PBL scheme in idealized TC simulations. The analysis focuses on two simulations, one that includes TKE advection (CTL) and one that does not (NoADV). Results show that relatively large TKE in the eyewall above 2 km is predominantly attributable to vertical advection of TKE in CTL. Interestingly, buoyancy production of TKE is negative in this region in both simulations; thus, buoyancy effects cannot explain observed columns of TKE in TC eyewalls. Both horizontal and vertical advection of TKE tends to reduce TKE and vertical viscosity in the nearsurface inflow layer, particularly in the eyewall of TCs. Results also show that the simulated TC in CTL has slightly stronger maximum winds, slightly smaller radius of maximum wind, and similar to 5% smaller radius of gale-force wind than in NoADV. These differences are consistent with absolute angular momentum being advected to smaller radii in CTL. Sensitivity simulations further reveal that the differences between CTL andNoADVare more attributable to vertical advection (rather than horizontal advection) of TKE. Recommendations for improvements of PBL schemes that use predictive equations for TKE are also discussed.