On the prediction of trajectory model error

A methodology for the development of an empirical, predictive model for forecasting trajectory model error is described. Trajectory model error is assumed to be a function of the complexity of the meteorological scenario under study. This complexity is quantified using a meteorological complexity factor (MCF), defined by the dispersion of a set of stochastic trajectories generated by random perturbations superimposed upon an observed wind field. Comparison between published trajectory model error data for specific cases and the corresponding MCFs would yield a means of forecasting trajectory model error. A weak relationship is found between MCFs and trajectory model errors for 22 cases, but a large degree of scatter precludes the development of a predictive model based on the MCF. Several explanations for this result are presented, including possible limitations in both the definition of MCFs as well as in the trajectory error data. Suggestions are given for further efforts to develop a predictive model for trajectory errors.