Impacts of assimilating various remotely sensed atmospheric parameters on WRF’s tropical cyclone prediction skill

Abstract:

Assimilating remotely sensed atmospheric parameters are critical for improving numerical weather prediction model skill, and especially for the prediction of tropical cyclone (TC) activities. The model skill is assessed by comparison with IBTRACs. In this talk, we will present results recently obtained using the weather research and forecasting data assimilation (WRF_DA) code. In the four TC cases studied (between 2003 and 2009), QuikSCAT measured near surface wind vectors (within a 6-hour assimilation window centered at model initiaisationl time) are assimilated. We further assimilated Infrared Atmospheric Sounding Interferometer (IASI) clear sky radiance and SSM/I measured total precipitable water vapour. By comparing with the control case (without assimilating any remote sensing data), the information content and impact of individual data sources are estimated. Possible use of cloudy and cloud contaminated radiances also will be assessed. Since the lifetime of a satellite platform is limited (~10 years), we further discuss a generic quality control scheme and an objective scheme of channel selection. This differs from the WRF_DA default procedure. An efficient method of obtaining bias correction coefficients are presented together with updating these coefficients in the prediction cycle.