Peter Childs and Sethu Raman Department of Marine, Earth, and Atmospheric Sciences North Carolina State University, Raleigh, NC 27695-8208

A rapid advance in computer processing speeds is allowing numerical weather simulations to be regularly generated at scales well less than synoptic. Assimilation systems coupled with numerous remotely sensed observations, standard surface and upper-air observations and aircraft observations are providing our numerical weather models a fairly accurate initial atmospheric state. However, few mesoscale modeling systems take full advantage of the semi-daily, high resolution scan of the sea surface temperature (SST) by our orbiting satellites. This study presents results from a model simulation for a spring 2003 case that was initialized with 1.4 km SST data. The main questions are: how does the real-time SST data influence the mesoscale model forecast along the coast and are the differences significant?

The study utilizes the MM5 mesoscale modeling system which has been modified to ingest the real-time sea surface temperature. The MM5 domain is centered over the North and South Carolina border with a base grid resolution of 15 km. Results from a nested 5 km grid, centered over the same area are assessed. For the case study two simulations are conducted, one with the 1.4 km sea temperature data and one with the typical coarse SST data from the 40 km NCEP ETA model, which is normally used to initialize the MM5. The simulation spans a period of several days when synoptic conditions allow for the local land and sea breeze circulations to dominate. The cases are analyzed for differences that are directly related to the sea surface temperature initialization. Additionally, surface and buoy observations are used to evaluate the performance of each simulation.

Initial results indicate that the better resolved SST boundary condition results in a more accurate model simulation, especially over an area like the Southeast US coast where spatial variations in the SST are large. Increased SST resolution also affects the strength of the sea breeze circulation. Properties of the land breeze circulation are also better modeled when the detailed SST is incorporated.