Hurricane Irene is the first major storm of the 2011 season, with a forecast path that is tracking up the US East Coast. The Weather Underground 2011 Hurricane Irene page is a good one page summary of data, and it will be accurate after this page is outdated.
Hurricane forecasting uses a series of models to estimate the position and intensity of storms. With a system this size, there is uncertainty as to the ultimate storm path, and it's not uncommon for a 5 day forecast to note that it might be more than 200 miles off course.
The National Hurricane Center's Irene archive has a fascinating loop of 5 day Hurricane Irene forecasts, showing the evolving best forecast track for the storm. It's a clear look not only at where the storm is going but also where projections have changed, in this case the projections that (on Tuesday) have landfall in North Carolina instead of Florida.
The Tropical Cyclone Guidance Project at NCAR has an even more detailed set of data for Hurricane Irene, including an archive of sets of model predictions for the storm.
After the storm, the set of models will be compared to the actual path to verify model accuracy. This National Weather Service National Hurricane Center Forecast Verification report charts forecast accuracy over time, noting how improved modeling and added computing power has made storm forecasting more accurate over time. How much better? Consider that the average 24 hour forecast error was 100 miles in 1990, but only 50 miles in 2010; also, that the models currently in use give as good a 5 day forecast (about 225 miles off) as the 2 day forecast was in 1988.
Edward Vielmetti's Wordpress weblog 