Using Observations to Improve Hurricane Initialization

Xiaolei Zhou
Florida State University

A bogus vortex must be constructed and used for initializing a hurricane forecast model if the hurricane sits over a data-void region. If the hurricane is not located over a data-void region due to the availability of high-resolution, abundant remote-sensing observations, is a bogus vortex still needed? The answer to this question is “yes” for the reasons stated below. Hurricane regions are areas where observations are often problematic due to rain contamination with QuikSCAT surface winds, cloud contamination with satellite radiances, the multi-path problem in the lower troposphere with GPS radio occultation measurements, etc. In addition, advances in data assimilation techniques and massively parallel interface computers now make it possible to carry out hurricane prediction at a high horizontal resolution of a few kilometers and vertical resolution less than a few hundred meters. However, the analyzed vortex is often too weak and misplaced to represent a vortex at such high resolutions. These conclusions suggest a need for a bogus vortex. The second question to be asked is then: Can a more realistic bogus vortex be constructed using satellite observations? The answer to this question is also “yes”. Traditionally, a bogus vortex is an artificial initial vortex specified based on a few available observational parameters, some empirically-assigned horizontal and vertical structures of a selected model variable, and several simplified dynamical constraints. Satellite remote-sensing observations provide direct measurements of various aspects of a real hurricane, such as tropopause distribution (TOMS ozone), surface vorticity (QuikSCAT surface wind), warm-core temperature (infrared radiance), hydrometeor information (microwave radiance), vertical variability (GPS radio occultation), etc. A more realistic bogus vortex could be constructed if this satellite observational information within a hurricane is incorporated into a vortex bogusing scheme. An effort has thus been made toward developing an improved hurricane initialization using various satellite observations and a 4D-Var vortex-bogusing method. A comparison between traditional vortex bogus schemes and a 4D-Var vortex-bogusing scheme will be made in terms of their advantages and disadvantages. Different considerations and preliminary results with applications of various conventional and satellite observations will be presented at the workshop.

Audio (MP3 File, Podcast Ready) Presentation (PowerPoint File)

Back to Small Scales and Extreme Events: The Hurricane