Process-oriented diagnostics: Principles, practice, community development, and common standards

However, the step from a diagnostic based on relationships among variables, even when hypothesis driven, to specific guidance for revising model formulation or parameterizations can be substantial. The POD may provide more informa-tion than a purely performance-based metric, but a gap between POD principles and providing actionable information for specific model revisions can remain.

Furthermore, in coordinating diagnostics development, there is a trade-off between freedom for the developer, aiming to capture innovation, and near-term utility to the modeling center. Best practices that allow for the former, while conforming to specifications that aid the latter, are important for community diagnostics development that leads to tangible model improvements. Promising directions to close the gap between principles and practice include the interaction of PODs with perturbed physics experiments and with more quantitative process models as well as the inclusion of personnel from modeling centers in diagnostics development groups for immediate feedback during climate model revisions.