Simulation of historical hurricane events using 20th Century Reanalysis

Simulation of historical events prior to the launch of weather balloons has suffered from the inability to produce high resolution mesoscale simulations given the lack of realistic, balanced, 3D fields for initial or boundary conditions for those model runs. Recently, however,. NOAA/CIRES have released the 20th Century Reanalysis Project Grids - - 3D grids of the atmosphere that are an estimate of that D structure arrived at using only the wealth of surface data available as far ago as the early 20th century (1908). Despite the lesser quality overall of such grids, they provide a homogeneous and global dataset that can be used for the first time to simulate historical events of interest.

The talk will present and discuss high resolution numerical simulations (using WRF and MM5) of several historical hurricane cases, initialized with this reanalysis, including:

1935 Florida Keys Major Hurricane 1938 New England Hurricane 1944 New England Hurricane 1954 Hurricane Carol 1954 Hurricane Edna

For cases such as the 1938 New England hurricane, where the structure of the storm at landfall was in question, the simulated storm will be examined within the cyclone phase space to given further insight into the possible structural evolution. The 1938 hurricane simulation was a surprisingly good simulation, with the landfall point 96hr in the future only about 50km off and with a reasonable intensity -- although the timing of the landfall was about 8hr off -- consistent with current biases for East Coast hurricanes.

The impact of a synthetic vortex on the initialization will be examined, and if the datasets are available, the ensemble members about the mean of the 20th century reanlaysis dataset will be used to produce a spread of potential simulated intensities, structures, and tracks.

The presentation will shed light on historical, yet less studied events, and to speculate on how well such events would be forecast today were they to occur again.