Due to the potential landfall of Hurricane Frances in the Southeastern United States, the National Hurricane Center (NHC) tasked a series of operational synoptic surveillance missions to gather dropwindsonde data in the tropical cyclone environment with the goal of improving numerical track forecasts upon which specialists rely for their official forecasts. A total of eight missions were conducted, starting on August 29, 2004 and ending just before landfall on the Florida east coast the early morning of September the fourth. The first two missions (August 29 and 30) were generally successful, with moderate reductions in forecast track error by the Global Forecast System (GFS). Aberson (2002) showed that in surveillance missions during 1997 and 1998, the amount of forecast improvement was proportional to the amount of aerial coverage of dropwindsonde data in the tropical cyclone environment. Since two planes were gathering data in the environment during the period in consideration, the expectation was for a forecast improvement.

However, the subsequent mission on September 1st resulted in increased track forecast errors up to 600 % throughout the entire forecast period. The synoptic situation is examined using station plots of the data obtained by the aircraft and GFS 500 hPa geopotential heights in both the operational and no dropwindsonde cases, to study the reasons for this failure. Additional cases for Hurricane Ophelia 2005 will also be presented.