Historic Rapid Intensification of Hurricane Patricia (2015): Cloud-Resolving Analysis and Prediction through Ensemble Assimilation of Dropsonde, Radar and Satellite Observations

In this study, we set out to capture and examine the extreme rapid intensification of Hurricane Patricia (2015), a storm that set many records including the intensification rate and intensity in terms of maximum observed wind speed, which was poorly forecasted by operational models. Using the Penn-State WRF-EnKF system assimilating all conventional observations and airborne Doppler radar observations and through sensitivity experiments, we demonstrate greatly improved analysis and intensity forecasts of Patricia with a 1-km horizontal grid spacing and a newly proposed surface drag parameterization of Chen and Yu (2016) in which C_d decreases for wind speeds greater than hurricane force. The 1-km simulations are critical to resolve the tight eyewall gradients of Patricia, and to well capture the rapid intensification of Hurricane Patricia following the assimilation of Doppler radar observations for the first P-3 aircraft flight. The need for high density and regular inner-core observation is seen and prediction skill is reduced between P-3 flights. Following the second P-3 flight the deterministic forecast again greatly improves and is able to capture the peak intensity of Patricia. EnKF Analysis is compared with observations from the ONR-sponsored TCI field campaign, including high density dropsondes within the inner-core of Patricia. We also plan to investigate the use of all-sky satellite observations to help fill this need for high density inner-core observations when in-situ aircraft observations are not available.

We further examine the predictability and dynamics of Patricia through convection-permitting forecasts initialized with the EnKF analysis perturbations at selected lead times. We seek to determine the inner-core dynamics versus environmental conditions necessary for both the rapid intensification and the record observed peak intensity. In addition, we examine the energetics and dynamics that result in the extraordinary peak intensity of Patricia. Our focus is on factors influencing the timing of rapid intensification and the peak intensity across the ensemble. Finally, model simulations are also compared with observations from the TCI field campaign, again including high density dropsondes within the inner-core of Patricia.

Figure 1: (top) Deterministic forecasts from 1-km cycling data assimilation for hurricane Patricia and (bottom) Deterministic and 10-member ensemble forecasts of Patricia from 2100 UTC 21 October, 2015.