Our coupled models, initialized at 8pm on September 7th, have completed and are now able to resolve Irma making landfall in Florida sometime between 2am and 5am Sunday morning. As a result, both the 3 and 7 day runs are capable of resolving the storm and currently point the landfall as occurring between Miami and Homestead (where Andrew made landfall in August 1992). This is in agreement with last night’s National Hurricane Center (NHC) forecast. The forecast track has shifted back slightly west since then, but is still within the NHC track forecast cone.
In addition to the 7-day animation shown, here is a KMZ file showing the predicted storm impact on southern Florida 54-hours into our forecast period (early Sunday morning). Included in the KMZ file are 10m wind + direction, sea level pressure (SLP) + wind direction, sea surface temperature (SST), significant wave height, estimated 1-hr precipitation, and simulated radar reflectivity for 06Z (2am EDT) Sunday morning. Load the files into Google Earth (available for free) to visualize the data.
We will continue to update as the storm progresses.