The Impact of HDOB and CIMSS AMV Observation Assimilation on the Analysis of Inner Core and Outflow Structures and the Prediction of Rapid Intensification of Hurricane Matthew (2016)

Hurricane rapid intensification (RI) prediction remains a big challenge in numerical weather prediction. For example, Hurricane Matthew (2016) underwent extreme RI, intensifying from a Category 1 storm to a Category 5 hurricane within 24 hours under a strong shear environment. However, most models failed to capture this RI. Early studies suggest accurate analysis of both the inner core and outflow structures are important for hurricane intensity forecast. However, limited or no inner core and outflow observations were assimilated in NWS operational HWRF model before the onset of RI for Matthew (2016), although both the high resolution HDOB and AMV data were collected at that time. The goals of the study are to (1) explore the best way to assimilate these data; (2) study the impact of assimilating these observations on the analysis of both the inner core and outflow structures; and (3) examine the impact of these data on the prediction of RI for Matthew.

Experiments are first conducted to investigate the optimal way of assimilating the HDOB observations (FL and SFMR observations, which are the only available inner-core observations before RI onset of Matthew) using the GSI based, continuously cycled hybrid DA system. It is found that a spin-down issue exists when HDOB data are assimilated by using the baseline 3DEnVar hybrid method. Further investigations show that this short-term spin-down is caused by a spurious secondary eyewall produced by the 3DEnVar analysis. With proper pre-processing of the HDOB observations and by using a 4DEnVar method, the inner-core structure analysis was improved. In addition, the RI prediction is more consistent with the best track without spin-down for the first 24 hours.

Assimilating CIMMS AMV observations on top of the HDOB observations further improve both the track and intensity forecasts. Specifically, both the magnitude and timing of the peak intensity are further improved.

Diagnostics are conducted to understand how the assimilation of these different types of observations impact RI prediction. During the first 18 hours, baseline experiment without assimilating the AMV data over-predict the RI. The assimilation of CIMMS AMV correctly weakens the upper-level outflow and therefore reduces the RI rate. In addition, the assimilation of FL and SFMR change the structure of the inner core vortex such as reducing the low-level moisture and therefore convection especially in the down shear left (DSL) region, which further correctly reduces the RI rate. These results together with further diagnostics will be presented in the conference.