Type of Document	Dissertation
Author	Park, Kyungjeen
Author's Email Address	kpark@met.fsu.edu
URN	etd-04092004-153121
Title	Incorporating TPC Observed Parameters And Quikscat Surface Wind Observations into Hurricane Initialization Using 4D-VAR Approaches
Degree	Doctor of Philosophy
Department	Meteorology, Department of
Advisory Committee	Advisor Name Title Xiaolei Zou Committee Chair Albert I. Barcilon Committee Member Ionel M. Navon Committee Member Peter S. Ray Committee Member T. N. Krishnamurti Committee Member
Keywords	Data assimilation 4D-Var Initialization Size Hurricane
Date of Defense	2004-03-30
Availability	unrestricted
Abstract This study aims to develop an objective hurricane initialization scheme which produces a model hurricane that not only satisfies forecast model constraints but also incorporates observed features such as the initial intensity (e.g., central sea-level pressure and maximum surface wind) and size (e.g., the radius of 34kt wind). It is based on the four-dimensional variational (4D-Var) bogus data assimilation (BDA) scheme originally proposed by Zou and Xiao (1999). The 4D-Var BDA scheme consists of two steps: (i) specifying a bogus sea level pressure (SLP) field based on parameters observed by the Tropical Prediction Center (TPC) and (ii) assimilating the bogus SLP field under a forecast model constraint adjusting all model variables. This research focuses on improving the specification of the bogus SLP indicated in the first step. First, an empirical linear model is developed for determining the size of the bogus vortex based on the TPC observed radius of 34kt. Numerical experiments are then carried out with and without making use of this model for the initialization and prediction of Hurricane Bonnie (1998) over the Atlantic Ocean. After runing these experimets, the track prediction is found to be less sensitive to bogus SLP than the intensity prediction. The maximum track error is less than 110 km during the entire three-day forecast using 4D-Var BDA generated initial vortices in both situation. However, the experiment using the linear model for the size specification considerably outperforms the other formulations in terms of the intensity prediction of Hurricane Bonnie. A further effort is made to incorporate QuikSCAT surface wind observations into the above hurricane initialization procedure, which aims to produce a more realistic initial vortex. Directly assimilating QuikSCAT surface winds results a negligible impact on hurricane prediction, so an indirect technique needs to be developed in order to maximize the impact of QuikSCAT observations on hurricane prediction. A radial profile of SLP is first derived from QuikSCAT surface winds, which is incorporated into the 4D-Var BDA procedure. Differences using the QuikSCAT-derived bogus SLP (QuikSCAT-BDA) and the above linear-model-derived SLP (LM-BDA) are examined for the prediction of Hurricane Gordon (2000). Compared with an HRD observed radial wind profile for Hurricane Gordon, the QuikSCAT-derived SLP profile is more realistic than the SLP profile specified by LM-BDA. However, numerical results show that LM-BDA performs better than QuikSCAT-BDA for both the track and intensity forecasts. Diagnosis of model output indicates that the LM-BDA generates larger surface fluxes of sensible heat, latent heat and moisture, as well as stronger downward angular momentum transport than the QuikSCAT and other BDA schemes do. These enhanced energy supplies offset the energy lost caused by friction and gravity wave propagation, allowing for the model to maintain a strong and realistic hurricane during the entire forward model integration.
Files	Filename       Size       Approximate Download Time (Hours:Minutes:Seconds)     28.8 Modem   56K Modem   ISDN (64 Kb)   ISDN (128 Kb)   Higher-speed Access    dissertation.pdf 5.79 Mb 00:26:48 00:13:47 00:12:03 00:06:01 00:00:30