Tropical Storm Nicholas (2003) formed over the eastern tropical Atlantic during mid October and moved generally northwest over the open ocean for approximately a week, then weakened to a tropical low on 23 October due primarily to increasing wind shear. Between 23 October and 1 Nov, the tropical low meandered across the Western Atlantic Ocean. On 1 Nov, the remnants of Tropical Depression Nicholas was moving southwestward towards the Florida peninsula as a shallow disorganized warm-core low that was in a transition stage to more of a cold core system. As the system approached and moved across Florida, an opportunity existed to examine the utility of a recently configured local mesoscale data assimilation and short-term prognostic system for providing additional guidance for forecasters.

The Advanced Regional Prediction System (ARPS) and the ARPS Data Analysis System (ADAS) were locally configured at NWS Melbourne to ingest a plethora of high resolution, non-traditional data sets, including observations from 44 Kennedy Space Center wind towers, five 915-MHz Doppler Radar Wind Profilers (DRWP), a 50-MHz DRWP, 30 Florida Automated Weather Network (FAWN), and a network of Automated Position Reporting System (APRS) WXNET sites. Other data sets include: Geostationary Operational Environmental Satellite (GOES-12) 1-km resolution visible and 4-km infrared imagery, METAR surface, buoy, C-MAN, ship reports, Aircraft Communications Addressing and Reporting System (ACARS) reports, and Melbourne (KMLB) Weather Surveillance Radar - 1988 Doppler (WSR-88D) base radial velocity and reflectivity fields. The ADAS domain covers most of Florida and was configured to produce analyses every 15-min with a 4-km horizontal resolution, using the 40-km Rapid Update Cycle (RUC) for initialization. Four ARPS cycles are run daily over the same domain and resolution as ADAS to produce 9-hr forecasts.

In this study, ARPS forecast fields and corresponding ADAS analyses were re-examined in a post analysis mode for the 3 November 0300 UTC forecast cycle as the tropical low approached central/south Florida. The impact of the high resolution local data assimilation was examined and compared to other available analysis and short-range forecast guidance (LAPS, MSAS, RUC). Fields such as surface moisture divergence were found to depict favorable areas for enhanced near-shore convection and the evolution of strong transport winds correlated well with gusty surface winds associated with showers. ADAS/ARPS also handled positioning of the low to mid level circulation of the tropical low better than traditional model analyses and RUC forecasts as the system traversed the state. Improved depiction of the circulation center and forecast fields helped to increase forecaster confidence of impending dangers to mariners and aviation concerns on a local scale. Although this system overall was poorly organized, local impacts were pronounced and ADAS/ARPS output with high resolution data assimilation demonstrated an ability to aid forecasters toward a positive enhancement of short-term forecasts.