地球惑星科学専攻 地球科学輻合部特別講演会

日時: 2010年1月22日(金) 15:00 - 16:30
場所: 京都大学理学部1号館 563号室
題目: Targeted Observation, Data Assimilation, and Tropical Cyclone Dynamics and Predictability
講演者: Dr. Chun-Chieh Wu (Department of Atmospheric Sciences, National Taiwan University, Taiwan)

講演要旨:
  In this lecture, highlights of three advanced tropical cyclones (TC) research issues are
presented: 1) the targeted observation of tropical cyclones (TC) in DOTSTAR (Dropwindsonde
Observations for Typhoon Surveillance near the Taiwan Region) and T-PARC (THORPEX -
Pacific Asian Regional Campaign) in 2008, 2) a newly-developed TC initialization and assimilation
method based on EnKF, and 3) a unique investigation of dynamical processes during eyewall
evolution in Typhoon Sinlaku during T-PARC.
  Three observation operators are designed for the TC assimilation, including the vortex center
location, vortex translation vector, and lower-level axisymmetric component of the wind-radius
profile.  It is shown that after the 24-h assimilation period in the high-resolution mesoscale model,
with the above three parameters assimilated through EnKF, a more realistic, dynamically well-balanced,
and model-consistent vortex structure can be constructed.  This method provides a useful means
to improve the initial condition for TC model simulations and predictions, especially when detailed
TC structure data are available.  The valuable TC data obtained from multiple airplanes in DOTSTAR
and T-PARC are then assimilated through the continuous update cycle to reconstruct a model-
observation consistent dataset, which can be used to examine the evolution of TC track, intensity
and structure change.
  Typhoon Sinlaku is a case in point under DOTSTAR and T-PARC with the most flight observation
with the potential to address major scientific issues in T-PARC ranging from structure change,
targeted observations to extra-tropical transition.  By assimilating all available dropwindsondes and
SFMR (Stepped Frequency Microwave Radiometer) data from 9 T-PARC flight missions with 157
dropwindsondes deployed for Typhoon Sinlaku (2008) during September 9 to 13, a unique high-spatial/
temporal-resolution and model/observation-consistent dataset is constructed for Sinlaku.  Further
sensitivity simulations of Sinlaku are conducted starting at different initial time, and therefore with
different data assimilated.  It is striking that some of the simulations are able to capture the secondary
eyewall cycle, while the others with fewer data assimilated are not. This provides an opportunity to
investigate the dynamical processes of the concentric eyewall formation in Sinlaku and its predictability.
  In particular, important parameters, including the filamentation time and beta skirt (vorticity gradient)
in 2D-turbulence context, are calculated to examine the validity of the extant theory on the formation
of concentric eyewall.  The uncertainty of the ensemble members in predicting the TC track, eyewall,
and structure evolution also provides useful physical insights into the TC predictability.

問い合わせ先: 石岡 圭一