Sensitivity to vertical shear and moist stability in a probabilistic regional dynamical model

Dynamical models are surpassing the skill of statistical models in seasonal prediction of tropical cyclone activity. This development makes it easier to explore the causes of inter-annual and longer-term variability. The modest success of statistical forecasts suggests that the causes can be understood as a one-way interaction in which the slowly evolving, large-scale tropical environment determines the frequency of cyclogenesis. This study builds upon the demonstrated success of a climate-downscaling procedure with an atmosphere-only model to gauge the importance of two key aspects of the seasonal-mean environment, the vertical shear and the thermodynamic stability.

Dynamical constraints on the imposed environment limit the range of practical perturbations available for sensitivity experiments. However, by making assumptions of linear sensitivity and by generating ensembles for each environment, one can reach a reasonably quantitative assessment of the separate importance of key environmental factors. Perturbations for this study are applied to the three-month (ASO) averages of the reanalysis fields.

Results show that the seasonal-mean thermodynamic stability plays an approximately equal role to the vertical shear in increasing storm activity in the Atlantic during 1980 to 2006. In the contrast between the extreme seasons 1982 and 1995, there appears to be a contribution from interannual differences at synoptic space and time scales along the eastern boundary. This makes attribution of the sharp interannual increase in activity more uncertain. However, a reasonable interpretation is that the seasonal-mean thermodynamic profile plays a lesser role than the shear in this case than in the multi-decadal trend.