Tropical cyclones (TCs) are among the most extreme and dangerous weather phenomena on Earth. In the United States, TCs that touch land cause an average of $10 billion in damage per year (Pielke et al. 2008). Hurricane Katrina alone (2005) caused $81 billion in damage and took more than 1,800 human lives. In developing countries, TC landfalls can be extremely harmful. For example, Cyclone Nargis (2008) caused the deaths of over 130,000 people in Myanmer (Burma). Due to the catastrophic nature of TCs, substantial efforts have been devoted to short-term predictions of TC pattern and intensity in an attempt to minimize damage and casualties. In recent years, the relationship between TC activity and climate change has attracted strong attention in the atmospheric research community. There is considerable evidence to suggest that regional CT activity is undergoing substantial change, which may be a response to climate change. As discussed later in the chapter, however, detecting and associating changes in TC activity resulting from climate change have been controversial due to uncertainties in the observed data and the difficulty in separating natural variability and anthropogenic forcing. Projections of TC's future activity are also not unanimous; Sources of uncertainty are still being identified and their relative effects tested in relation to unanimous projections of TC activity. However, the need for reliable projections for future TC activities is strong among government organizations and industries located near coasts that are often affected by TC landfalls. Therefore, identify the source of uncertainties in future projections of TC activity and develop a mechanism to reduce the paper bill by half when evaluating simulated TC climatology. Furthermore, explanations outside the TC detection methodology (such as large-scale simulated conditions) are preferred to support the simulated TC climatology. Therefore, the specific objectives of this work are:1. Quantify the range of sensitivity to various CT tracking criteria (Chapter II);2. Identify model biases independent of CT tracking sensitivity (Chapter III), as well as new sources of uncertainties and intrinsic limitations so far undocumented (Chapters III and IV);3. Increase the robustness of future projection of CT activity with respect to sensitivity to CT monitoring (Chapter IV); e4. Present an alternative approach to represent the limitation of CT intensity projection using a statistical model (Chapter V). A summary and further implications can be found in Chapter VI.