We use the CM1 model to simulate tropical cyclones in an idealized framework, with horizontal grid spacing as fine as 30 meters. At this grid spacing, the scales of the vortices are well resolved. To help us interpret our dataset of observed extreme updrafts and wind speeds, we examine millions of virtual dropsondes (modified parcel trajectories that have a fall speed relative to the air) within our simulations. Dropsonde observations of extreme updrafts are often characterized by large vertical gradients in both kinematic and thermodynamic variables. Using the virtual sondes, we show that this apparent vertical variability is likely an artifact of the sondes being horizontally advected through dynamical features, and that the true structure of these features is quite different than can be inferred from the dropsondes.
The strongest updrafts and horizontal wind speeds in the LES are somewhat stronger than what we have previously observed in intense tropical cyclones. By sampling the virtual sondes in a realistic manner, we show that a typical virtual flight mission substantially underestimates the true peak gust. It is very unlikely for a given virtual sonde to sample wind speeds exceeding the current observational maximum (111 m/s), even though 120-130 m/s winds are present somewhere within the simulated TC at nearly all times. The statistical distributions of wind speed and updraft magnitudes in the LES compare well with observations, which suggests that the strongest wind gusts within real tropical cyclones may indeed be on the order of 120 m/s or more.