Extended Abstract (1.3M)3.3
Mesoscale FDDA Experiments with ACARS Data
PAPER WITHDRAWN
Chia-bo Chang, Texas Tech University, lubbock, TX, Texas; and R. Dumais
This is a Four Dimensional Data Assimilation (FDDA) study using MM5 in conjunction with high-frequency ACARS (Aircraft Communications Addressing and Reporting System) data collected by commercial aircraft during both en route and ascent/descent phases of their flights. ACARS wind and temperature observations recorded at about every ten minutes crisscrossing the nation are ideal for testing the effects of FDDA on short-term mesoscale numerical weather prediction (NWP). The research is conducted based on real-data simulation experiments. The model domain used is 1320 by 1320 km2 covering bulk of Texas and the grid size is 20 km. There are 24 levels in the vertical. The initial state and lateral boundary conditions are derived from the Eta model analyses. The method of observational nudging is employed in FDDA.
A severe weather event in central Texas was selected for this study. Ninety five ACARS profiles were used. We had conducted a large number of 36-h single-grid MM5 runs including sensitivity tests of model performance to the nudging parameters. Without FDDA MM5 performed well in the middle layers but under-predicted temperatures near the surface. The model QPF patterns compared well with the observations. FDDA had considerable impact throughout the entire model atmosphere. The influence of ACARS data was still quite significant at 12 h after the termination of FDDA. The ACARS data distribution was not very uniform in the lower layers. Some rather large changes at the surface were found far away from the areas where ACARS observations were concentrated. Also, there were substantial changes in convective precipitation and QPF patterns in some areas. The domain means of various difference fields between the runs with and without ACARS data showed some temporal oscillations as the model was undergoing adjustment to the data insertions. The amplitudes were much larger in the upper layers for both winds and temperatures, while the frequencies varied with variables and layers. The phases of oscillations did not show any clear correlation between the upper and lower layers or any coherent relationships between variables. The sensitivity tests revealed that shortening the length of nudging or reducing the size of ACARS data set did not result in smaller amplitudes. The experiments suggest that ACARS data has potential to significantly enhance the mesoscale NWP expertise. We believe more case studies of diverse mesoscale events (e.g., oscillation drylines, rainstorms, wind storms) are needed. We would gain a physical understanding of how to capitalize the high resolution and frequency data for a wide range of mesoscale weather and develop real time guidance in applying this valuable data source.
Supplementary URL: http://www.chiabo.ttu.edu
Session 3, Numerical weather prediction, data assimilation
Monday, 6 August 2007, 3:30 PM-5:30 PM, Waterville Room
Previous paper Next paper