Abstract
This dissertation aims to improve tropical cyclone (TC) intensity forecasts by exploring the connection between intensity forecast error and parameters representing initial condition uncertainty, atmospheric flow stability, TC strength, and the large-scale environment surrounding a TC. After assessing which of these parameters have robust relationships with error, a set of predictors are selected to develop a priori estimates of intensity forecast accuracy for Atlantic basin TCs. The applications of these forecasts are then discussed, including a multimodel ensemble that unequally weights different intensity models according to the situation. The ultimate goal is to produce skillful forecasts of TC intensity error and use their output to enhance intensity forecasts.