Internally versus Externally Forced Tropical Cyclone Variability

Seasonal predictions of tropical cyclone activity using dynamical models are being generated at ever increasing resolution, yet the confidence we should place on any one future prediction is largely unknown. It has been shown that the internal variability, defined here as the variance of an ensemble of predictions differing only in their initial condition, increases in importance as the size of the region of interest decreases. Internal variability may be particularly acute for convective precipitating systems such as tropical cyclones due to stochastic contributions to their occurrence.

Over what spatial and temporal scales does internal variability dominate tropical cyclone activity and does internal variability impose an upper limit on seasonal and decadal predictability? An initial condition ensemble experiment is presented to determine the relative importance of internally vs. externally forced variability of North Atlantic tropical cyclone activity. Results suggest simulated internal variability dominates on at least annual and perhaps on longer timescales. The sources of internal variability will be discussed, including easterly waves and mesoscale convective processes. The implications of the findings for seasonal and decadal prediction will be discussed.