Weather and climate phenomena develop as part of the coupled ocean-atmosphere-land-ice system. To understand the nature of the coupled system and its constituent processes, as well as the basis for and the limits of their predictability, some important concepts are reviewed, including determinism, chaotic error growth, and linear as well as nonlinear perturbation dynamics. It is shown that weather is predictable but only for finite times. Initial and model errors amplify, eventually rendering weather and climate forecasts useless. First, skill is lost in forecasts of fine-scale features while larger-scale phenomena remain predictable for longer periods of time. Processes and other characteristics associated with different scales of motion are discussed next, proceeding from the finest to the largest, coupled global-scale ocean-atmosphere phenomena. The potential use of ensemble forecast techniques to quantify scale and case-dependent predictability in the context of hydrologic forecasting is emphasized throughout.