The position uncertainty of tropical cyclones (TCs) may stretch along specific directions, yet the global pattern of the dominant direction of uncertainty remains unclear. This study proposes an analogue ensemble algorithm based on historical best-track data to quantify TC track uncertainties in the along-track (AT) and cross-track (CT) directions. The results reveal a strong latitudinal variation in the dominant direction of uncertainty: AT uncertainty is larger near the equator and within the mid-latitude westerlies, whereas CT uncertainty increases in recurvature zones. Moreover, the ratio of CT to AT uncertainty exhibits a negative correlation with TC translation speed. A comparable anisotropic pattern of track uncertainty is also identified in multi-model ensemble forecasts, indicating that the state-dependent preferred direction of position variability is a natural characteristic of TC track predictions. These findings provide the global-scale quantification of intrinsic anisotropic TC track uncertainty and highlight the importance of representing anisotropic uncertainty in probabilistic risk assessment and operational forecasting.

Read original article

Read original article