The southwestward motion and unexpected landfall of Hurricane Katrina over South Florida caused 14 deaths in Florida and an estimated $700 million worth in damages. In order to reduce the loss of life and damage to critical infrastructure in future hurricane forecasts, it is imperative to conduct reanalysis studies to identify flaws in current forecasts. This study reviews the National Hurricane Center's (NHC) advisories and the Hurricane Rainband Intensity Change Experiment (RAINEX) group's weather discussions and numerical model forecasts for Katrina from its genesis to its southwestern track over South Florida. First, the differences between the forecast track errors from the genesis of this storm until Hurricane Katrina emerged in the Gulf of Mexico were compared with the forecast track errors of this storm from the Gulf of Mexico to Louisiana to demonstrate that the first landfall made by this storm was much more poorly predicted than the storm's second landfall. The rawinsonde and dropsonde data (obtained from the NOAA G-IV aircraft) were then compared with the NCEP Reanalysis. It was found that an elongated high pressure system over Texas had split, forcing Katrina to move towards the southwest. The comparisons demonstrated this by showing two distinct centers of two High pressure centers on 24 August 2005 8PM EDT which was visible in both the NCEP Reanalysis and the observational data sets. Finally, the difference fields between two high-resolution MM5 model forecasts based on initial and boundary conditions from the Environmental Canada Global model (CMC) and the NCEP Global Forecasting System (GFS) respectively were compared. The respective predictions of the primary features that influenced Katrina's path were contrasted.