Many environmental processes evolve over space and time, creating a complex dynamical system.The construction of nonlinear regression models that describe the space-time evolution of complex processes will be useful in many applications.
Statistical modeling of dynamical systems makes the estimation and construction of confidence intervals for interesting quantities from data possible. This involves fitting nonlinear models and estimating dynamical systems quantities of interest such as global and local Lyapunov exponents. The evolution of cloud cover over time and its space-time relationship to other climate variables is another interesting dynamical system and very important in climate modeling. We present the results of a neural network to model the nonlinear nearest-neighbor grid cell relationships over time. We develop a stability analysis for quantifying the predictability of space-time processes. Space-time Lyapunov exponents are based on extending the dynamical system quantity that quantifies the short-term growth of a perturbation in time, to a space-time interpretation.
|