A Feasibility Study on Improving Tropical-Cyclone Inner-Core Structure Based on Data Assimilation of Inner-Core Atmospheric Motion Vectors

Abstract

The present operational estimation of the tropical cyclone (TC) intensity is made by subjectively combining observational methods including the Dvorak technique. However, as the kinds and frequencies of satellite observations are increased, it is desirable to switch to physics-based objective analysis. This study is conducted to test its feasibility by examining the assimilation of high-resolution atmospheric motion vectors (AMVs) obtained by a recently proposed method. Even though the method can retrieve low-level high-speed winds near the inner edge of the eyewall cloud, it cannot directly measure the maximum wind speed, which is normally achieved in the eyewall. On the basis of an ensemble Kalman filter assimilation system, observing system simulation experiments (OSSEs) are performed by assimilating pseudo-AMV observations created from a nature run during the mature stage in a real TC. Mimicking the real AMVs, the pseudo-observations are limited to low-level AMVs in the eye without the maximum wind in the eyewall. The AMV assimilation limited to the eye reduces the spread of the TC-center position among the ensemble members. Although the present OSSE has a limitation to assume that the model used is unbiased and highly dense AMVs are always obtained in a vertical layer of the eye, these results suggest that assimilating high-resolution low-level AMVs in the eye can be a hopeful approach to realize improvement of the TC structure with small errors in the model for the physics-based objective analysis of TC intensity in the future.