Numerical Simulations of Landfalls of Hurricane Sandy (2012) and Rita (2005): Sensitivity to Initial Conditions

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner
Wednesday, 7 January 2015
Zhaoxia Pu, University of Utah, Salt Lake City, UT; and S. Zhang

Accurate forecasts of landfalling hurricanes are of great importance, yet forecast challenges are always presented. In this study, we investigate the impact of initial conditions on the accuracy of the prediction of Hurricane Sandy (2012) and Rita (2005) during their landfalls. Various initial and boundary conditions derived from ECMWF analysis, NCEP Global Forecast System (GFS) final analysis (FNL), and North American Model (NAM) analysis, are used in the numerical simulations with an advanced research version of the weather research and forecasting (WRF ARW) model. Different forecast lead times are also examined. The model simulations are validated by observations made from the NASA TRMM satellite data, radar and dropsonde observations collected by the NOAA Hurricane Research Division (HRD) operational/research flights, as well as ground-based radar observations.

It is found that the numerical simulations of Hurricane Sandy (2012) and Rita (2005) are sensitive to the initial conditions and forecast lead time. Discrepancies are found in various experiments in terms of hurricane landfall times and locations. The intensity and structures of the hurricanes also vary in different experiments.

Further diagnoses are conducted to compare the key atmospheric variables in the varied initial conditions in order to investigate the reasons that lead to the large discrepancies in the forecasts. In addition, for some cases, results are also compared with those simulations generated from the Hurricane WRF (HWRF) model in order to obtain additional insight on the effects of model uncertainties when compared with the impact of initial uncertainties in prediction of landfalling hurricanes. Finally, data assimilation experiments are performed to confirm major findings from this study.