Abstract
Using the model twin formulation, the predictability of the zonal index, an inherently low-frequency feature of the circulation, is studied in a two-layer, global, zonally homogeneous model. While typical periods for variability in the zonal index are on the order of 100 days, the index is predictable with useful skill only out to 23 days. The predictability of the zonal index shares some behavior with the predictability of time averages. A simple model of the growth of forecast error in the zonal index is constructed. The model demonstrates that the combination of a long timescale with relatively poor predictability is consistent with the behavior of a feature that acts as a dynamical filter of synoptic scale noise.
© 1992 Walter A. Robinson, Jianchun Qin, published by Stockholm University Press
This work is licensed under the Creative Commons Attribution 4.0 License.
