The objective in this project is to employ numerical integration of simple systems of difference or differential equations to investigate the following issues, (1) the rate at which errors, or differences between realizations, grow during the initial stages of prediction and the rate at which they approach their limiting values during the final stages, (2) the amount of predictive information that may be added by combining past states with the present state, and (3) the size of a region of phase space in which predictions are approximately linear, so that "analogue" forecasting is feasible. This study is important because it will help to improve our understanding of the behavior of dynamical systems, such as the atmosphere, and their prediction.
