Recorded presentation8A.2
Synoptic sensitivity analysis of Typhoon Sinlaku (2008) and Hurricane Ike (2008)
William Komaromi, Univ. of Miami/RSMAS, Miami, FL; and E. D. Rappin, S. J. Majumdar, M. J. Brennan, S. G. Chen, D. S. Nolan, R. Langland, and C. S. Velden
The track forecasts of the landfall of Hurricane Ike and Typhoon Sinlaku (both 2008) presented significant challenges to the forecaster, due to uncertainty in the global models. One method to improve the skill and remove uncertainty in NWP of tropical cyclones is to assimilate high-resolution data such as rapid-scan wind vectors aboard geostationary satellites, in ‘target' areas in which the forecast is deemed to benefit optimally from the data. Prior to the actual targeting of these data (which involve quality control and data assimilation), it is instructive to understand directly the dominant initial-condition sensitivity in the synoptic environment.
In order to achieve this, a vorticity perturbation method (using inversion techniques) is employed within a 27 km resolution WRF-ARW modeling framework. The selection of features in which to perturb the vorticity field is based on a combination of subjective analysis, objective targeting methods, and null hypotheses. For example, local positive and negative vorticity perturbations were created within a shortwave trough situated to the northwest of Sinlaku, over Russia. Alternatively, an Ensemble Transform Kalman Filter (ETKF) sensitivity analysis based on the ECMWF ensemble indicated that Sinlaku would be more sensitive to inaccuracies associated with the region downstream of the trough. These results will be compared against evaluations of NOGAPS simulations that use additional satellite wind data.
For Hurricane Ike, potential sensitive areas include Tropical Storm Lowell in the eastern Pacific, a pair of shortwaves situated near Alaska, and a large trough over the Great Lakes. For the latter case, the reduction in 850-250 hPa vorticity within the trough led to a more westward 4-day forecast of Ike's landfall than that produced by the unperturbed simulation. This and other sensitivity results and diagnostics will be presented for both tropical cyclones.
Session 8A, T-PARC Targeted Data Studies
Wednesday, 12 May 2010, 8:00 AM-9:45 AM, Arizona Ballroom 6
Previous paper Next paper