Multi-parameter sensitivities along warm conveyor belt trajectories: A visual analysis using Met.3D

Abstract

Cloud microphysical processes are highly relevant for cloud and precipitation characteristics, cloud radiative properties and the latent heat release during phase changes of water can interact with atmospheric dynamics. These sub-grid scale processes are typically parameterized in numerical weather prediction models, introducing parametric uncertainty in weather forecasts. The analysis of uncertainties related to these parameterizations imposes multiple challenges: On the one hand, it requires robust quantification of the impact of hundreds of uncertain model parameters. On the other hand, it requires adequate tools to filter, visualize, and understand the parameter impacts. Algorithmic Differentiation (AD) is a tool to efficiently evaluate the magnitude and timing at which a model state is sensitive to a model parameter [1]. We demonstrate the capabilities of AD, focusing on uncertain parameters in a two-moment cloud microphysics scheme along trajectories of a warm conveyor belt, which is the primary cloud- and precipitation-forming airstream in extratropical cyclones. To understand the parameter influence, we here introduce methods to systematically analyze different impacts in different warm conveyor belt ascent scenarios [2]. For example, this includes an objective clustering of trajectories w.r.t to parameter sensitivities. Met.3D, an open-source tool for interactive, three-dimensional visualization of numerical atmospheric model datasets, then provides a visual interface to compare multiple sensitivities on multiple trajectories from each cluster, assess the spatio-temporal relationships between the sensitivities and the trajectories’ shapes and locations, and find similarities in the temporal development of sensitivities along various trajectories’ location and time for ascent.    [1] Hieronymus, M., Baumgartner, M., Miltenberger, A. and Brinkmann, A.: Algorithmic Differentiation for Sensitivity Analysis in Cloud Microphysics, J. Adv. Model Earth Syst. (2022), 10.1029/2021MS002849.  [2] Neuhauser, C., Hieronymus, M., Kern, M., Rautenhaus, M., Oertel, A., and Westermann, R.: Visual analysis of model parameter sensitivities along warm conveyor belt trajectories using Met.3D (1.6.0-multivar0), Geosci. Model Dev. Discuss. [preprint], https://doi.org/10.5194/gmd-2023-27, in review, 2023.