Abstract
The flow dependence of atmospheric predictability implies that forecast errors grow more rapidly in some atmospheric conditions than in others. A better understanding of this flow dependence thus requires a local analysis of error growth. To facilitate such an analysis, this study introduces a feature-based perspective. While feature identification and tracking is often applied to atmospheric systems, associated forecast errors exhibit smaller-scale structure and thus lack spatial coherence. Consequently, using a standard feature approach, merging and splitting of features is ubiquitous, which severely limits the ability to automatically identify distinct temporal feature evolutions and subject them to statistical analysis. While spatial filtering of data alleviates this inherent challenge, it does not resolve it and incurs a loss of information. We overcome this challenge by introducing a feature post-processing that combines individual features into regional-scale entities, which exhibit much increased spatial and temporal coherence. It is these post-processed entities that prove suitable for subsequent feature-based analysis. We demonstrate the utility of the feature-based perspective by applying it to the spread of global ensemble experiments designed to assess upscale error growth. Analyses are exemplified that contribute to an improved understanding of the flow dependence of error-growth mechanisms and that link error growth characteristics to local atmospheric conditions.
© 2025 American Meteorological Society. This is an Author Accepted Manuscript distributed under the terms of the default AMS reuse license. For information regarding reuse and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).
Jorge de Heuvel’s current affiliation: Humanoid Robots Lab, University of Bonn, Bonn, Germany
