Prediction and predictability of Hurricane Joaquin (2015) revealed through convection-permitting real-time deterministic and ensemble forecasts

Based on successful real-time deterministic and ensemble forecasts of Hurricane Joaquin (2015) using the experimental WRF-EnKF hurricane analysis and prediction system operated at the Pennsylvania State University (PSU), this study examines the predictability and dynamics of Joaquin with a focus on its deterministic track prediction uncertainty.

We systematically compare the deterministic forecasts provided by the PSU WRF-EnKF experimental system with the operational HWRF running at NCEP; HWRF persistently predicted the storm to make landfall along the US east coast during the early stages of the storm. We find the difference between the PSU WRF-EnKF and NCEP HWRF is mainly caused by the difference in initial conditions of the incipient core vortex instead of the large-scale environmental or differences in the forecast models. The HWRF forecasts initialized with PSU EnKF initial conditions correctly capture the southern loop and its subsequent off-coast track.

Additionally, the PSU WRF-EnKF ensemble demonstrates extreme sensitivity of Joaquin's track forecasts to small differences in the initial conditions. Nearly half of the 60 ensemble members initialized with the EnKF perturbations at 1200 UTC 29 September make landfall along the US east coast while the other half correctly trend out-to-sea. The extreme track sensitivity is also reflected in the large divergence among different operational forecast models suggesting a low predictability of Joaquin early in its lifecycle.

Finally, we examine the deterministic and ensemble forecasts of the heavy precipitation associated with Joaquin that led to record-breaking flooding in the south east coast of the United States, which was predicted well by the PSU WRF-EnKF real-time deterministic forecasts at lead times up to 5 days. The influence of Joaquin versus midlatitude system on the heavy precipitation will be explored through extensive ensemble sensitivity analysis in particular by comparing members with different tropical cyclone tracks or with different amount of precipitation.