Recent observing system experiments (Uhlhorn and Nolan, 2012) suggest that current technology (e. g. SFMR) sampling of a numerical model "nature run" hurricane could miss (underestimate) small convective scale wind maxima by an average of 7%. However such models are very sensitive to parameterizations and tuning coefficients, and are not capable of simulating atmospheric turbulence (Rotunno and Bryan 2012, Bryan 2012, Rotunno et al., 2009) so they are missing details seen in real hurricanes. In real storms such wind maxima would likely be sampled by Doppler radars aboard NOAA's P3 research aircraft, so undersampling should not be an issue if NOAA aircraft are monitoring the storm. Alternatively, if the reconnaissance aircraft is not equipped with Doppler radar and the SFMR manages to sample a small scale convective feature, the associated wind maximum could be interpreted as a transient feature unrepresentative of storm intensity (Brennan et al., 2009, Powell et al., 2011). Uhlhorn and Nolan's work showed that today's sampling technology is capable of resolving the 10 min mean wind field of the hurricane to a high degree of accuracy. We believe that the mean wind field is a more reasonable metric to observe, monitor, and predict than the current definition of intensity. AOML is currently exploring the analysis and predictability of the mean wind, and the concept of estimating intensity and gusts as statistical properties based on the mean wind (Marks and Vukicevic 2012).

References