Probabilistic analysis of the dynamics and predictability of tropical cyclogenesis

This study first examines the dynamics and predictability of a tropical disturbance to the east of the incipient tropical cyclone Alex in the Atlantic basin during the 2004 hurricane season. Both the disturbance preceding Alex and the present disturbance were encompassed by similar environments, yet only the Alex disturbance developed. However, at least one mesoscale deterministic model and its stochastic ensembles develop the non-forming system into a mature tropical cyclone in as few as 72 hours.

Taking advantage of discrepancies between ensemble members, statistical correlation is used to elucidate why storms in some ensemble members strengthen drastically quicker than others. Also investigated are the impacts of convective parameterization and of slight changes to the initial ensemble mean. When convective parameterization is used, it is found that higher MUCAPE, more abundant deep moisture, and lower static stability at the PBL height combine together to produce more active initial convection and stronger storms. When convective parameterization is not used, high MUCAPE and deep moisture seem to be the most important factors needed for more active convection and stronger storms. Regardless of differences between ensembles, in this case it is clear that an atmosphere supporting vigorous deep convection ultimately favors cyclogenesis and that the extreme sensitivity of cyclogenesis to moist convection causes large ensemble spread.

Despite the results above, differences in initial MUCAPE only explain about 25-50% of the variation within the ensembles, and these ensembles exhibited some fairly large changes in cyclone intensity given only slight variations in the initial variables. For example, a difference in initial conditions as little as 1/4 of the initial difference between some ensemble members can make the difference between a non-developing disturbance (through 36-h) and a storm near hurricane strength.

Thus, future work includes investigating individual ensemble members to see what exactly governs cyclogenesis versus non-formation when differences in initial conditions are minute. The new work will entail tracing cyclogenesis in several ensemble members to see how ensemble members evolve in light of the previous results.

Similar work on a different developing tropical cyclone (Allison, 2001) reveals that system strength therein is also strongly tied to early precipitation intensity and coverage. However, in the case of Allison, initial CAPE seemed to matter little; rather it was the strength of the low-level jet and moisture surge to the south of where Allison developed in the model that seemed to most strongly govern cyclogenesis.