26th Conference on Hurricanes and Tropical Meteorology

Thursday, 6 May 2004: 4:00 PM
MM5 Simulations of Precipitation and Mesocyclone Dynamics Associated wih TC Gabrielle (2001) using High Resolution Data of East Central Florida
Napoleon II Room (Deauville Beach Resort)
G. V. Rao, Saint Louis University, St. Louis, MO; and K. Santhanam, J. W. Scheck, S. M. Spratt, B. C. Hagemeyer, R. Edwards, J. Schaefer, J. L. Case, and E. Kemp
Poster PDF (204.0 kB)
MM5 Simulations of Precipitation and Mesocyclone Dynamics Associated with TC Gabrielle (2001) Using High Resolution Data of East Central Florida

Gandikota V. Rao, Krishnaraj Santhanam and Joshua W. Scheck Department of Earth and Atmospheric Sciences, Saint Louis University, St.Louis, MO

Scott M. Spratt and Bartlett C. Hagemeyer NOAA/National Weather Service, Melbourne, FL

Roger Edwards and Joseph Schaefer NOAA/NWS Storm Prediction Center, Norman, OK

Jonathan L. Case ENSCO, Inc., Cocoa Beach, and Applied Meteorology Unit, NASA, Kennedy Space Center,FL

Eric Kemp Center for Analysis and Prediction of Storms, University of Oklahoma, Norman, OK

Abstract

Tropical Cyclone (TC) Gabrielle formed as a tropical depression in the Gulf of Mexico on 11 September, 2001 and made landfall as a tropical storm along the west coast of Florida between Tampa Bay and Fort Myers on 14 September 2001. Convection associated with an outer rain band of TC Gabrielle rotated onshore of east central Florida about 1000 UTC 14 September. Convection cells within the rain band developed mesocyclones, which were evident in the National Weather Service Melbourne, Florida Doppler radar imagery. The mesocyclones spawned three confirmed tornadoes between 1155 and 1255 UTC resulting in property damage estimated at $500,000.

In this study, MM5 simulations are performed using local data assimilation from an extensive collection of network sites to examine whether the kinematic fields such as low-level vorticity could serve as proxy for the radar-observed mesocyclones. If such indications are present, then, the dynamics of the simulated mesocyclones can be better understood resulting in greater situational awareness and forecaster confidence in advance of TC tornado events and potential issuance of considerably longer tornado watch and warning lead times.

NCEP Final Analyses (FNL) were obtained from NCAR and used to initialize the MM5 integrations. Four domains were used, with grids that were 2-way interacting. The mother domain covered most of the United States, at a 54-km resolution. Grids of 18-km, 6-km and 2-km were nested to resolve successively finer scales, the last domain covering a high resolution data rich area over east central Florida. High-resolution ADAS analyses (Advanced Regional Prediction System-Data Analysis System) encompassing KMLB Level II reflectivity and radial velocity, surface METAR observations, ACARS (Aeronautical Radio), Florida Automated Weather Network observations, Kennedy Space Center (KSC) / Cape Canaveral Air Force Station network of 44 wind and temperature sensors, one 50-MHz DRWP at KSC and 915-MHz DRWPs (profilers). These data were separately analyzed by ADAS using a 4-km resolution at 15-minute intervals.

Intermittent Data Assimilation (IDA) was then performed for every 15 minutes in an experimental mode in the 2-km MM5 domain from 1000 UTC to 1200 UTC on 14 September. Sensitivity runs were made by running MM5 with and without the high-resolution local data sets within the ADAS analyses. Tests were also performed to assess the sensitivity of the forecast fields using different boundary layer and cumulus parameterization schemes.

Preliminary results show the special data affects other domains within 4 hours. The location of the 12-hour predicted TC center appears to be related to the particular parameterization scheme. A 24-h simulation of accumulated (convective and non-convective) rainfall within the 18-km domain, beginning 0000 UTC 14 September 2001 and using the high-resolution data over east central Florida, seemed to agree well with the observed precipitation over Florida. The runs without the high-resolution data underestimated the rainfall, suggesting that supplemental high-resolution data can be utilized to achieve more accurate mesoscale forecast guidance.

The research was sponsored by a COMET Partners Project 5-00820

Supplementary URL: