Any study of atmospheric predictability must draw upon the vast body of important contributions of Ed Lorenz, this study will be no exception. This work will re-examine the question of local predictability which was brought to the fore by Lorenz and Emanuel in the context of the utility of adaptive observations. This study approaches the questions of local predictability and expected forecast skill using the tools of two formalisms which heretofore have not been widely incorporated into such problems: wave propagation and information theory.

Because the questions of adaptive observations are most naturally posed in physical space and not in a modal phase space, wave propagation ideas are the most convenient context within which to ask such questions as: Where should one observe the atmosphere today to increase our confidence in a 3 day forecast over San Diego? An interpretation of the adaptive observing problem using the concepts of complex phase and group velocities of amplifying baroclinic waves will be given. Additionally, it will be asserted that the relative information metric is the most useful measure for posing and gauging the efficacy of adaptive observing strategies. Examples of the usefulness of these concepts will be given, drawn from short-range ensemble forecasts using the NCAR CAM3.