The United States Weather Research Program has mandated the improvement of landfalling tropical cyclone (TC) forecasts. Current numerical weather prediction analysis and integration systems charged with predicting TC intensity and track (GFDL, Eta, and NOGAPS) demonstrate significant forecast skill. Little has been done, however, to assess nor improve forecast skill of predicted asymmetric wind and precipitation structures. This is true, in part, because conventional data sources necessary to assimilate and evaluate storm structure are absent over the open ocean where these storms develop. This research seeks to improve the forecast simulations of storm structure and track through the assimilation of non-conventional satellite and aircraft observations.
As an attempt to reduce current TC numerical forecast system inadequacies over the ocean, the University of Wisconsin (UW), Florida State University (FSU), and the National Centers of Environmental Prediction (NCEP) have initiated a collaborative effort to develop an integrated meso-beta analysis and model forecast system. The primary goal is to determine which analysis methodologies will most effectively utilize the vast numbers of remotely sensed data available from geostationary and polar satellites in addition to data sets derived from aircraft, drifting surface platforms, etc. The data assimilation framework will be NCEP's 3D-VAR working in conjunction with the cloud resolving University of Wisconsin Nonhydrostatic Modeling System (UW-NMS) and an axisymmetric hurricane simulation system also based on the UW-NMS. A major focus of the research will be the assimilation of the multi-level winds derived from GOES 8/9 infrared, water vapor, and visible radiance measurements as well as GOES 3-layer precipitable water data. The inclusion of the remotely sensed meso-beta scale momentum information particularly the high density water vapor wind vectors will decrease the reliance upon synthetic tropical cyclone observations (axisymmetric bogusing) and increase the quantitative spatial description of the hurricane vortex. It is expected that many of the satellite derived wind assimilation techniques already developed for the large scale will be scalable to a meso-beta assimilation cycle, though issues of observation representativeness may lead to some potential obstacles. The marriage of the axisymmetric vortex with satellite derived information as well as aircraft observations is expected to lead to improvements in assimilated storm structure.
In order to evaluate the impact of the assimilated data both before and after landfall, passive microwave measurements from Special Sensor Microwave Imagery (SSM/I) and combined passive microwave-precipitation radar measurements from the Tropical Rainfall Measuring Mission (TRMM) will be used with physically based precipitation profile retrieval algorithms operated at FSU. Beyond the evaluation of the assimilation of meso-beta scale information, the TRMM and SSM/I information may eventually be assimilated into the UW/NCEP system using a diabatic initialization scheme similar to that currently being developed for NCEP's Aviation model.
To test the initial inclusion of multi-level satellite derived winds, sensitivity experiments will be performed utilizing < 10 km horizontal resolution with the new integrated analysis and model forecast system for a particular tropical cyclone, hurricane Opal. Hurricane Opal is of particular interest because of its robust asymmetric structure, rapid intensification, and poor prediction by the operational models. The results of this case study, a complete description of the integrated UW/NCEP prediction system, and future endeavors will be presented.