A new method to simulate tropical cyclone boundary layers

Four so-called "simple" methods to simulate the boundary layer of tropical cyclones (TCs) are evaluated using dropsonde data and flight-level "flux run" observations. The modeling methods do not simulate an entire tropical cyclone (which can be computationally expensive). Rather, they simulate small areas, of order 5 km x 5 km, and account for the larger-scale tropical cyclone circulation through imposed pressure-gradient and inertial-type terms. The simplest method, analogous to the Ekman boundary layer approach, tends to over-predict TC inflow-layer depth by a factor of 2, and under-predicts wind speeds near the surface. Two recently proposed methods, which attempt to account for inertial effects in TCs, have the opposite problem: they under-predict inflow-layer depth by a factor of 2, and over-predict near-surface wind speed. A new method that accounts for radial advection of angular momentum in a TC produces inflow-layer depths and near-surface windspeeds that match observations well. Analysis of the latter method shows that turbulent momentum flux profiles in the TC boundary layer are not linear with height, and explains why certain aircraft "flux run" data tend to underestimate surface drag coefficient.